In 2015, 46% of the world's new HIV infections came from Eastern and Southern Africa, and nearly 40% of the infections in this region originated from South Africa alone.1 The “treat-all” policy recommended by the World Health Organization2 enables all people living with HIV access to antiretroviral therapy (ART) and offers promise for rapid reduction in HIV incidence in the region. This is because, in addition to the life-expectancy benefits of ART, ART has also been shown to greatly reduce the transmission risk from those infected.3 The key to optimizing the HIV-prevention benefits from ART is initiation of ART early in disease progression, which can only be achieved with early HIV diagnosis. Achieving early HIV diagnosis by ensuring regular HIV counselling and testing of those at risk is therefore critical to the reduction of incidence in the region.
South Africa has made considerable progress in increasing access to HIV testing services (HTS) in the country. The 2010 national HIV testing campaign conducted approximately 13 million HIV tests, with an estimated 7.6 million first-time testers between 2010 and 2012.4 In 2012, 65.5% of South Africans reported having ever-tested for HIV, an increase from 21.4% in 2002 and 50.8% in 2008.5 Furthermore, recent estimates based on mathematical modeling indicate that as much as 86% of people living with HIV in the country have been diagnosed.6 This estimate suggests that the country is close to achieving the first of the 90-90-90 targets set out by UNAIDS, which proposes that by 2020 coverage of HIV diagnosis should meet or exceed 90% of HIV-positive individuals.7 By contrast, however, a recent representative household survey in a hyperendemic region of South Africa8 indicated a significant shortfall from the UNAIDS 90-90-90 targets—only 65% of women and 52% of men who were HIV-positive knew their status.9 Shortfalls in reaching the “first 90” target have also been observed in 2016 in Malawi, Zimbabwe, and Zambia.10–12 The findings from these studies support previous findings that men and youth (aged 15–24) continue to lag behind in terms of knowledge of HIV status and contribute disproportionately to the shortfall in the “first 90.”4,13,14
Access to HTS in South Africa is primarily facility-based through public clinics and hospitals.5 This approach limits access to HTS to individuals seeking care, leaving many asymptomatic HIV-positive persons undiagnosed. Moreover, self-perceived risk of HIV infection has been shown to be highly inaccurate among individuals in South Africa.15 Consequently, large proportions of people living with HIV will be unaware of the need for HIV testing until they become ill. Alternative testing strategies that facilitate regular testing among asymptomatic individuals are therefore needed.
Home-based HTS (HBHTS) has been shown to increase the uptake of HIV testing in several African countries16 and to be effective in reaching men, young adults, and first-time testers.13 In South Africa, HBHTS has been shown to be less costly than clinic-based testing17 and effective at increasing testing uptake in rural areas.18 Moreover, with facilitated interventions, it has been shown that individuals diagnosed HIV-positive during HBHTS can be linked to care effectively.19 However, HBHTS also needs to be effective in reaching HIV-positive individuals who are unaware of their status. To date, little is known about HBHTS uptake among undiagnosed HIV-positive individuals. There are also limited data on the impact of HBHTS in enhancing progress toward the UNAIDS 90-90-90 targets and the role of HBHTS on awareness of HIV-positive status in regions with high coverage of individuals ever-tested for HIV.
This study measured the uptake of HBHTS and the impact of the home-test on knowledge of HIV-positive status among individuals in a hyperendemic area of South Africa. We examined the sociodemographic, psychosocial, and behavioral factors associated with HBHTS uptake, with a particular emphasis on understanding HBHTS uptake among undiagnosed HIV-positive men and women in high prevalence settings.
The data for this study were collected during the second cross-sectional survey of the HIV Incidence Provincial Surveillance System (HIPSS) conducted in the Vulindlela and Greater Edendale areas of the uMgungundlovu district of KwaZulu-Natal, South Africa. The area is predominantly rural to periurban, with limited employment opportunities and high levels of poverty. HIV prevalence is estimated to be one of the highest in the country. HIV prevalence among individuals aged 15–49 years in 2014–2015 was estimated to be 36%,9 and district-level antenatal prevalence in 2013 was 42%.20
Study Design and Procedures
The HIPSS design, described in detail elsewhere,8 involves serial cross-sectional household surveys with imbedded HIV incidence cohorts. The survey analysed in this study was conducted between July 2015 and June 2016 (referred to as the “second baseline” survey) and included 1 individual per household between the ages of 15 and 49 years. Multistage sampling was used to select a representative sample of individuals for the survey, with enumerated areas being the primary sampling unit.
The key objective of HIPSS was to determine HIV prevalence and incidence in the area. To this end, eligibility criteria for study enrollment were a willingness to provide peripheral blood samples for laboratory HIV testing. Thus, data on HIV status were available for all participants. Although all participants provided blood for laboratory HIV testing, results from these tests were only available to participants at local clinics after some time (should they have wished to access the results). All participants were however provided with the option of HBHTS to determine their status during the survey.
All participants provided written informed consent, peripheral blood samples for laboratory measurements (HIV, CD4 cell count, and viral load), and completed a standardized, face-to-face questionnaire capturing demographic, socioeconomic, and behavioral data. Self-reported data regarding HIV testing history, reasons for never testing, and perceived HIV status (based on self-reported results from participants' most recent HIV test: positive, negative, indeterminate, or did not respond) were captured as part of the questionnaire.
At the end of the study visit, all participants (regardless of self-reported HIV status) were offered HBHTS. Field staff members were certified to perform the HBHTS procedures in line with National Department of Health guidelines. The testing algorithm, described in detail elsewhere,21 involved a serial algorithm with 2 rapid diagnostic tests using blood specimens from finger pricks.22 The first rapid diagnostic test was performed with Alere Determine HIV-1/2 (Alere, Matsudo, Japan) and the second, if required, with UniGold HIV (Trinity Biotech, Bray, Ireland). Reasons for refusing the home-test were captured. The accuracy of the rapid tests used in the HIPSS study is described in detail elsewhere.21
Laboratory HIV testing was conducted using fourth generation HIV enzyme immunoassays to test for HIV antibodies and antigens. Positive HIV tests were confirmed with the HIV 1/2 Combi Roche Elecys (Germany) (Roche Diagnostics, Penzberg, Germany) and HIV-1 Western Blot Biorad assay (Bio-Rad Laboratories, Redmond, WA).
HIPSS was approved by the University of KwaZulu-Natal Biomedical Research Ethics Committee (BF269/13), in collaboration with the Department of Health, Province of KwaZulu-Natal (HRKM 08/14). The project was reviewed according to Centers for Disease Control and Prevention (CDC) human research protection procedures and was determined to be research, but CDC was not engaged.
Sampling weights were calculated to account for the unequal probability of selection associated with the multistage sampling design and nonresponse. The weights were benchmarked against the 2011 Census data by gender and 5-year age group to ensure that the weights of the participants sampled added up to the population estimates for the area. All statistics were estimated by applying survey weights and survey procedures that corrected for sampling and participation biases, and took into account clustering within enumeration areas. All analyses were performed using SAS (SAS Institute, Cary, NC, version 9.4 and STATA 13 (Stata Corporation LP, College Station, TX).
Descriptive statistics were reported for the full sample (n = 10,236) and for the sample of individuals who had never tested or did not report a positive result from their last HIV test (n = 7484), as interest in HBHTS among individuals who reported a positive result from their last HIV test was (as expected) minimal. Statistics were also reported for the sample of individuals who were HIV-positive (as determined by laboratory HIV test results) but undiagnosed before being offered HBHTS (n = 1118). People living with HIV were assumed to be undiagnosed before being offered HBHTS if they had a positive test result from the HIV laboratory test but reported never having had an HIV test or did not report a positive result from their last HIV test.
The impact of HBHTS on the proportion ever-tested for HIV was assessed by estimating the proportion of individuals who received their first HIV test during the HBHTS. The impact of HBHTS on knowledge of HIV-positive status was assessed by estimating the proportion of HIV-positive individuals (as determined by laboratory testing) who were made aware of their HIV-positive status through the HBHTS offered during the HIPSS survey.
Uptake of HBHTS was estimated by sex, age, and testing history. Log-binomial regression was used to assess the association between HBHTS uptake and a range of sociodemographic, behavioral, and health variables using risk ratios (RR). Separate exploratory models were run for each independent variable, selected on the basis that the factor could theoretically influence demand for HIV testing. Adjusted risk ratios (aRR) presented were adjusted to control for key demographic covariates, identified before analysis, that are associated with HIV testing: age group, highest education level, and relationship status.4,14,23,24 Models were estimated separately by sex.
Of the 17,199 eligible households approached during the second HIPSS baseline survey, 12,247 (71.2%) households agreed to participate. A total of 10,236 (83.6%) individuals were enrolled from these households; refusal rates of men and women were 19.5% and 13.1%, respectively. Of those enrolled, 48.2% were men and 51.8% were women (Table 1). The median age was 26 years for men and 27 years for women. HIV prevalence was 24.5% [95% confidence interval (CI): 23.0% to 26.1%] among men and 45.0% (95% CI: 43.4% to 46.7%) among women. Among all HIV-positive individuals, before the offer of HBHTS, 62.9% of men and 73.4% of women were correctly aware of their status as confirmed by laboratory results. The majority reported having had an HIV test at least once previously, with only 18.0% of men and 8.6% of women having never tested. The most common reasons for never having tested were being afraid to know one's status (men: 41.6% and women: 29.9%), and never previously considering an HIV test (men: 35.1% and women: 39.2%).
Among individuals who had never tested or did not report a positive result from their last HIV test, 76.1% of men and 85.5% of women reported that their last HIV test was negative. A further 0.3% reported that their last test result was indeterminate, 17.4% had never tested for HIV, and 1.0% refused to respond to the question regarding HIV status. Within this group, 10.8% of men and 17.9% of women tested HIV-positive through laboratory testing of peripheral blood samples.
Among undiagnosed HIV-positive individuals—those with a positive HIV laboratory test who did not self-report a positive result from their last HIV test—66.5% of men and 83.7% of women self-reported their HIV status to be negative. Approximately one quarter of undiagnosed HIV-positive men had never had an HIV test before the survey, with 48.1% of these men reporting fear, and 32.1% reporting never thinking of it as reasons for never testing. A total of 31.8% and 44.5% of undiagnosed HIV-positive men and women, respectively, reported that their last HIV-negative test result was received within the previous 12 months.
Impact of HBHTS
After HBHTS, the proportion of individuals ever-tested for HIV increased from 82.1% (95% CI: 80.5% to 83.6%) to 90.8% (95% CI: 89.8% to 91.8%) among men and from 91.4% (95% CI: 90.5% to 92.3%) to 96.2% (95% CI: 95.6% to 96.8%) among women (Fig. 1). The largest impact was observed in the 15–19 year age group, with HBHTS increasing the proportion ever-tested from 73.9% (95% CI: 70.4% to 77.4%) to 88.2% (95% CI: 85.7% to 90.7%) among men and from 70.5% (95% CI: 67.3% to 73.6%) to 88.0% (95% CI: 85.8% to 90.3%) among women.
Through HBHTS, the percentage of HIV-positive individuals who knew their status increased from 62.9% (95% CI: 59.4% to 66.4%) to 74.2% (95% CI: 70.8% to 77.5%) among men, and from 73.4% (95% CI: 71.5% to 75.4%) to 80.5% (95% CI: 78.6% to 82.4%) among women (Fig. 2). The greatest gains in knowledge of HIV infection were observed in the 15–19 year and 20–24 year age groups. The percentage of people living with HIV aged 15–24 years who knew their status increased from 36.6% (95% CI: 26.5% to 46.8%) to 59.3% (95% CI: 48.7% to 69.9%) for men, and from 50.8% (95% CI: 45.6% to 55.9%) to 64.8% (95% CI: 59.4% to 70.2%) for women.
Uptake of HBHTS
HBHTS uptake among individuals who had tested and reported a positive result from their last HIV test was only 0.4%. Among the subsample of participants who had never tested or did not report a positive result from their last HIV test, uptake of HBHTS was 48.1% among men and 52.6% among women (Table 2). HBHTS was accepted by 49.1% of men and 55.8% of women who had never tested before. Among individuals who reported a recent HIV test (within 12 months preceding the survey), 44.9% of men and 49.6% of women accepted the home-test.
Uptake of HBHTS was lower among the subsample of undiagnosed people living with HIV. In this group, 33% of men and 30.8% of women accepted testing. HBHTS was lowest among undiagnosed HIV-positive individuals who reported having had a test in the year preceding the survey (men: 24.3% and women: 21.6%). Uptake was highest among individuals who had previously never tested, particularly in the subsamples of people who had not tested because of a lack of time (men: 34.7% and women: 68.8%) and among those who had not thought of testing before (men: 56.9% and women: 55.4%).
Reasons for Refusing HBHTS
Among individuals who had never tested or did not report a positive result from their last HIV test, the most common reason provided for not accepting HBHTS was “I know my status” (men: 44.8% and women: 59.6%) (see Table S1, Supplemental Digital Content 1, http://links.lww.com/QAI/B239). Fear of knowing one's status was the second most common reason given for not home-testing, with 34.5% of men and 23.1% of women refusing to test because of fear. Among undiagnosed HIV-positive individuals, the most common reasons provided for refusing HBHTS were also “I know my status” (men: 27.7% and women: 45.4%) and fear (men: 45.1% and women: 32.9%).
Factors Associated With HBHTS Uptake
Among individuals who had never tested or did not report a positive result from their last HIV test (Table 3), HBHTS uptake was positively associated with perceived HIV risk: individuals were more likely to take up HBHTS if they perceived themselves at higher risk of acquiring the disease (men: aRR: 1.21; 95% CI: 1.11 to 1.32; women: aRR: 1.13; 95% CI: 1.04 to 1.23). HIV-positive individuals (as determined by the laboratory tests) were less likely to home-test compared with HIV-negative individuals (men: aRR: 0.66; 95% CI: 0.55 to 0.80; women: aRR: 0.54; 95% CI: 0.47 to 0.61). In addition, men who home-tested were more likely to have reported having a partner with a sexually transmitted infection (STI) than those who did not home-test (aRR:1.40; 95% CI: 1.16 to 1.68), and women who home-tested were more likely to report never using a condom in their last 3 relationships (aRR:1.10; 95% CI: 1.01 to 1.20).
Factors Associated With Uptake of HBHTS Among Undiagnosed HIV-Positive Individuals
Among undiagnosed HIV-positive men and women (see Table S2, Supplemental Digital Content 2, http://links.lww.com/QAI/B239 for regression results), there was a negative relationship between recent HIV testing and HBHTS uptake. Individuals were more likely to home-test if their last HIV test had been more than 12 months before (men: aRR:1.32; 95% CI: 0.83 to 2.11; women aRR:1.76; 95% CI: 1.32 to 2.35) or if they had never tested (men: aRR:1.61; 95% CI: 1.01 to 2.56; women aRR:1.96; 95% CI: 1.33 to 2.88). Those who accepted HBHTS were more likely to perceive themselves at risk of HIV acquisition in the future (men: aRR:1.46; 95% CI: 1.01 to 2.10; women aRR:1.47; 95% CI: 1.11 to 1.94). Among women, those who home-tested were more likely to report having had emotional support from a family member in the 12 months preceding the survey (aRR:1.37; 95% CI: 1.04 to 1.80). The CD4 count distribution of the home-testers was not significantly different to that of the nontesters.
In a high HIV burden region in South Africa, with adult (15–49 years) HIV prevalence of 24.5% among men and 45% among women, our study findings showed that most individuals had previously been tested for HIV (82.0% of men and 91.4% of women). However, despite the high coverage of HIV testing, the percentage of people living with HIV who knew their status fell considerably short of the UNAIDS target of diagnosing 90% of people living with HIV by 2020: 62.9% of HIV-positive men and 73.4% of HIV-positive women were aware of their status at study enrollment. The proportion of diagnosed HIV-positive individuals was particularly low among young (15–24 years) men (36.6%) and women (50.8%). These sex and age differentials in awareness of HIV status are consistent with results from nationally representative samples in South Africa25 and reflective of similar differentials in the uptake of HIV testing nationally.4 Our results indicate that the current UNAIDS estimate of 86% of South Africans living with HIV knowing their status6 does not necessarily apply to locations with high HIV burden.
The study results demonstrate that HBHTS can lead to substantial gains in the proportion of HIV-positive individuals aware of their status even in a context with high coverage of past HIV testing. Gains were greatest among younger populations, and particularly among young men, with a 62% relative increase in the proportion of 15–24 year -old HIV-positive men aware of their status. In addition, results indicate that HBHTS could potentially facilitate regular testing, as approximately 47.5% of our sample who had tested for HIV in the 12 months preceding the survey accepted HBHTS. The importance of repeat testing in regions with high HIV incidence is underscored by the study finding that almost half (44.5%) of undiagnosed HIV-positive women in the sample had been tested for HIV in the 12 months preceding the survey.
Patterns of uptake of HBHTS in our study also indicate the potential of this testing approach in reaching groups with historically poorer uptake of HIV testing: men, younger individuals, and healthier individuals.4,14,26 Our findings showed that uptake of HBHTS was similar by sex, age groups, and health status (as indicated by CD4 count). Moreover, indicative of the importance of taking HIV testing to certain individuals, approximately half the sample who had never previously tested also accepted HBHTS. HBHTS provided first-time testing to most individuals who reported never testing previously because of a lack of time or because they had not considered it, and to almost half the sample of people who had never tested previously because of fear of the diagnosis.
However, patterns of testing uptake revealed potential barriers to the efficacy of HBHTS. Approximately 67% of undiagnosed HIV-positive men and 69.2% of undiagnosed HIV-positive women refused HBHTS. Our findings suggest that refusal of HBHTS among undiagnosed individuals was linked to an incorrect perception of status and risk. Undiagnosed HIV-positive men and women were less likely to accept HBHTS if they did not perceive themselves at risk of HIV infection, and significant proportions of undiagnosed HIV-positive men and women who refused HBHTS reported knowing their status as a reason for refusal. Likely related to this phenomenon, HBHTS uptake was lower among undiagnosed HIV-positive individuals who reported testing for HIV in the 12 months preceding the survey; fewer than a quarter of these individuals agreed to the test. Finally, the data from this study suggest that fear continues to be an important deterrent to testing, particularly among undiagnosed HIV-positive men.
Our results should be considered alongside the study's limitations. Uptake of HBHTS was lower in this study than that observed in other studies in sub-Saharan Africa.13,16 The high coverage of past HIV testing observed in this study relative to other studies is one potential explanation. It is also possible that participants might have been deterred from accepting the home-test because of the time they had already spent on other aspects of the study. However, given that fewer than 8% of men and 3% of women reported a lack of time as a reason for refusing the HBHTS, we believe the impact of this would have been minimal. HBHTS in general offer testing at a household level. It is unclear what impact the HIPSS study design—the selection of 1 individual per household—had on uptake and the applicability of our results to HBHTS programs. Testing families as a unit has been endorsed as a positive benefit of HBHTS,27 and hence, it is possible that the lack of support from other family members may have undermined testing uptake. Self-reporting bias might have also influenced the accuracy of study measures. There is potential, for example, for underreporting of previous HIV-positive diagnosis.25 HIV prevalence estimates might also have been underestimated because of selection effects. Population-based surveys often have high non-participation in HIV testing, resulting in underestimates of HIV prevalence because of a positive correlation between non-participation and knowledge of HIV-positive status.26,28–30 We believe the potential for such bias is limited in our study because all study participants provided blood samples for laboratory HIV testing, and individual non-response in the study was low. High study participation may relate to the collection of peripheral blood being used to screen for several diseases, not solely HIV. Nevertheless, it is possible that large sex differences in HIV prevalence estimates in our study are related to differences in non-participation, and correlated factors, by sex.30 A study based on 2007 Zambia Demographic and Health Survey data, for example found that HIV prevalence estimates from models corrected for selection were substantially higher for men, but not women.30 It is possible that bias of this nature led to study results overestimating the impact of HBHTS on awareness of status among HIV-positive men. Finally, it is unclear how our findings would apply in other settings, such as regions with lower HIV prevalence.
Overall, results from our study underscore the importance of regular testing in regions with high HIV incidence, as is recommended by the World Health Organization.31 Our findings highlight the importance of community-based testing strategies, such as HBHTS, in increasing HIV status awareness among people living with HIV who would otherwise delay testing because they do not perceive the need to test, or are afraid to find out their status. HBHTS combined with interventions to facilitate linkage to care19 could play a major role in reducing high rates of HIV infection among young men and women in the area. Our results also suggest that HBHTS programs need to consider and counter significant barriers to uptake of testing among undiagnosed HIV-positive individuals to maximize the effect of HBHTS on progress toward the UNAIDS 90-90-90. Opt-out HBHTS or incentivized testing32,33 may be useful strategies to increase uptake of testing among undiagnosed HIV-positive individuals who do not perceive the need for testing. HBHTS should also be designed to encourage testing among individuals who tend to refuse testing due to fears around knowing their HIV status. Future research is required to determine the optimal frequency of HBHTS in regions with high HIV incidence.
The authors thank all the study participants, study staff, coinvestigators from Epicentre, CAPRISA, HEARD, NICD, and CDC and district primary health care clinic staff. The authors thank our collaborating partners: The National Department of Health, Provincial KwaZulu-Natal Department of Health, uMgungundlovu Health District, the uMgungundlovu District AIDS Council, HIV and AIDS/STI/TB (HAST) unit KwaZulu-Natal, local municipal and traditional leaders, and community members for all their support throughout the study.
1. UNAIDS. Prevention Gap Report. Geneva, Switzerland: UNAIDS; 2016.
2. WHO. Consolidated Guidelines on the Use of Antiretroviral Drugs for Treating and Preventing HIV Infection. Geneva, Switzerland: WHO; 2016.
3. Cohen MS, Chen YQ, McCauley M, et al. Prevention of HIV-1 infection with early antiretroviral therapy. N Engl J Med. 2011;365:493–505.
4. Maughan-Brown B, Lloyd N, Bor J, et al. Changes in self-reported HIV testing during South Africa's 2010/2011 national testing campaign: gains and shortfalls. J Int AIDS Soc. 2016;19:20658.
5. Shisana O, Rehle T, Simbayi L, et al. South African National HIV Prevalence, Incidence and Behaviour Survey, 2012. Cape Town, South Africa: HSRC Press; 2014.
6. UNAIDS. Ending AIDS: Progress towards the 90-90-90 Targets. Geneva, Switzerland: UNAIDS; 2017.
7. UNAIDS. 90-90-90: An Ambitious Treatment Target to Help End the AIDS Epidemic. Geneva, Switzerland: UNAIDS; 2014.
8. Kharsany AB, Cawood C, Khanyile D, et al. Strengthening HIV surveillance in the antiretroviral therapy era: rationale and design of a longitudinal study to monitor HIV prevalence and incidence in the uMgungundlovu District, KwaZulu-Natal, South Africa. BMC Public Health. 2015;15:1149.
9. Grobler A, Cawood C, Khanyile D, et al. Progress of UNAIDS 90-90-90 targets in a district in KwaZulu-Natal, South Africa, with high HIV burden, in the HIPSS study: a household-based complex multilevel community survey. Lancet HIV. 2017;4:e505–e513.
10. Ministry of Health, Malawi. Malawi Population-based HIV Impact Assessment (MPHIA) 2015-16: First Report. Lilongwe, Malawi: Ministry of Health, Malawi; 2017.
11. Ministry of Health, Zambia. Zambia Population-based HIV Impact Assessment (ZAMPHIA) 2016: First Report. Lusaka, Zambia: Ministry of Health, Zambia; 2017.
12. Ministry of Health and Child Care, Zimbabwe. Zimbabwe Population-based HIV Impact Assessment (ZIMPHIA) 2015–16: First Report. Harare, Zimbabwe: Ministry of Health and Child Care (MOHCC), Zimbabwe; 2017.
13. Sharma M, Ying R, Tarr G, et al. Systematic review and meta-analysis of community and facility-based HIV testing to address linkage to care gaps in sub-Saharan Africa. Nature. 2015;528:S77–S85.
14. Huerga H, Van Cutsem G, Ben Farhat J, et al. Who needs to be targeted for HIV testing and treatment in KwaZulu-Natal? Results from a population-based survey. J Acquir Immune Defic Syndr. 2016;73:411–418.
15. Maughan-Brown B, Venkataramani AS. Accuracy and determinants of perceived HIV risk among young women in South Africa. BMC Public Health. 2017;18:42.
16. Sabapathy K, Van den Bergh R, Fidler S, et al. Uptake of home-based voluntary HIV testing in sub-Saharan Africa: a systematic review and meta-analysis. PLoS Med. 2012;9:e1001351.
17. Tabana H, Nkonki L, Hongoro C, et al. A cost-effectiveness analysis of a home-based HIV counselling and testing intervention versus the standard (facility based) HIV testing strategy in rural South Africa. PLoS One. 2015;10:e0135048.
18. Doherty T, Tabana H, Jackson D, et al. Effect of home based HIV counselling and testing intervention in rural South Africa: cluster randomised trial. BMJ. 2013;346:f3481.
19. Naik R, Doherty T, Jackson D, et al. Linkage to care following a home-based HIV counselling and testing intervention in rural South Africa. J Int AIDS Soc. 2015;18:19843.
20. Department of Health, South Africa. The National Antenatal Sentinel HIV Prevalence Survey, South Africa, 2013. Pretoria, South Africa: Department of Health; 2014.
21. Kufa T, Kharsany AB, Cawood C, et al. Misdiagnosis of HIV infection during a South African community-based survey: implications for rapid HIV testing. J Int AIDS Soc. 2017;20(suppl 6):21753.
22. Department of Health, South Africa. National HIV Counselling and Testing Policy Guidelines South Africa. Pretoria, South Africa: Department of Health; 2015.
23. Hensen B, Lewis JJ, Schaap A, et al. Factors associated with HIV-testing and acceptance of an offer of home-based testing by men in rural Zambia. AIDS Behav. 2015;19:492–504.
24. Venkatesh KK, Madiba P, De Bruyn G, et al. Who gets tested for HIV in a South African urban township? Implications for test and treat and gender-based prevention interventions. J Acquir Immune Defic Syndr. 2011;56:151–165.
25. Johnson LF, Rehle TM, Jooste S, et al. Rates of HIV testing and diagnosis in South Africa: successes and challenges. AIDS. 2015;29:1401–1409.
26. Sharma M, Barnabas RV, Celum C. Community-based strategies to strengthen men's engagement in the HIV care cascade in sub-Saharan Africa. PLoS Med. 2017;14:e1002262.
27. Mantell JE, DiCarlo AL, Remien RH, et al. “There's no place like home”: perceptions of home-based HIV testing in Lesotho. Health Educ Res. 2014;29:456–469.
28. McGovern ME, Marra G, Radice R, et al. Adjusting HIV prevalence estimates for non-participation: an application to demographic surveillance. J Int AIDS Soc. 2015;18:19954.
29. Hogan DR, Salomon JA, Canning D, et al. National HIV prevalence estimates for sub-Saharan Africa: controlling selection bias with Heckman-type selection models. Sex Transm Infect. 2012;88(suppl 2):i17–i23.
30. Bärnighausen TBJ, Bor J, Wandira-Kazibwe S, et al. Correcting HIV prevalence estimates for survey nonparticipation using Heckman-type selection models. Epidemiology. 2011;22:27–35.
31. WHO. Consolidated Guidelines on HIV Testing Services. Geneva, Switzerland: WHO; 2015.
32. Nglazi MD, van Schaik N, Kranzer K, et al. An incentivized HIV counseling and testing program targeting hard-to-reach unemployed men in Cape Town, South Africa. J Acquir Immune Defic Syndr. 2012;59:e28–e34.
33. Kranzer K, Govindasamy D, van Schaik N, et al. Incentivized recruitment of a population sample to a mobile HIV testing service increases the yield of newly diagnosed cases, including those in need of antiretroviral therapy. HIV Med. 2012;13:132–137.