HIV self‐testing: breaking the barriers to uptake of testing among men and adolescents in sub‐Saharan Africa, experiences from STAR demonstration projects in Malawi, Zambia and Zimbabwe

Abstract Introduction Social, structural and systems barriers inhibit uptake of HIV testing. HIV self‐testing (HIVST) has shown promising uptake by otherwise underserved priority groups including men, young people and first‐time testers. Here, we use characteristics of HIVST kit recipients to investigate delivery to these priority groups during HIVST scale‐up in three African countries. Methods Kit distributors collected individual‐level age, sex and testing history from all clients. These data were aggregated and analysed by country (Malawi, Zambia and Zimbabwe) for five distribution models: local community‐based distributor (CBD: door‐to‐door, street and local venues), workplace distribution (WD), integration into HIV testing services (IHTS), or public health facilities (IPHF) and during demand creation for voluntary male medical circumcision (VMMC). Used kits were collected and re‐read from CBD and IHTS recipients. Results Between May 2015 and July 2017, 628,705 HIVST kits were distributed in Malawi (172,830), Zambia (190,787) and Zimbabwe (265,091). Community‐based models, the first to be established, accounted for 519,658 (82.7%) of kits distributed, with 275,419 (53.0%) used kits returned. Subsequent model diversification delivered 54,453 (8.7%) test‐kits through IHTS, 23,561 (3.7%) through VMMC, 21,183 (3.4%) through IPHF and 9850 (1.7%) through WD. Men took 294,508 (48.2%) kits, and 263,073 (43.1%) went to young people (16 to 24 years). A higher proportion of male self‐testers (65,577; 22.3%) were first‐time testers than women (54,096; 17.1%) with this apparent in Zimbabwe (16.2% vs. 11.4%), Zambia (25.4% vs. 17.7%) and Malawi (27.9% vs. 25.9%). The highest proportions of first‐time testers were in young (16 to 24 years) and older (>50 years) men (country‐ranges: 18.7% to 35.9% and 13.8% to 26.8% respectively). Most IHTS clients opted for HIVST in preference to standard HTS in each of 12 delivery sites, with those selecting HIVST having lower HIV prevalence, potentially due to self‐selection. Conclusions HIVST delivered at scale using several different models reached a high proportion of men, young people and first‐time testers in Malawi, Zambia and Zimbabwe, some of whom may not have tested otherwise. As men and young people have limited uptake under standard facility‐and community‐based HIV testing, innovative male‐ and youth‐sensitive approaches like HIVST may be essential to reaching UNAIDS fast‐track targets for 2020.

Low HIV testing, knowledge of status, and suboptimal treatment and prevention coverage among men and young people (15 to 24 years) in sub-Saharan Africa are key gaps in the HIV response. Recent population-based HIV impact assessments (PHIA) in Zimbabwe, Malawi and Zambia, showed that men with HIV were less likely to know their status than HIVpositive women [5][6][7]. Less than half of youth aged 15 to 24 years with HIV knew their status, which was substantially lower than coverage in older age groups [5][6][7]. Demographic and health surveys (DHS), conducted in 30 sub-Saharan African countries during 2011 to 2016, showed lower testing coverage among men compared to women for all age groups except 45-to 49-year olds [3].
Low coverage (defined as the proportion of population eligible for an intervention that has received it) of HIV testing and treatment among men in Africa is often due to poor utilization of public sector health facilities, reflecting both social and structural health systems barriers [8,9]. Prevailing social norms around masculinity that emphasize toughness, self-reliance and sexual success lead to an avoidance of health services, among other consequences [10][11][12][13]. For HIV, this is compounded by anticipated loss of social standing and sexual desirability if diagnosed HIV-positive, increasing fear of stigma and promoting a mindset whereby testing when "still healthy" is considered undesirable [10,13]. Greater formal and informal employment among men compared to women, can also hinder access due to job insecurity and high opportunity and indirect costs [13]. Likewise, young people have well-described agespecific barriers that make use of existing facility-based HIV testing services especially difficult [14]. Recognizing and responding with innovative male-and youth-sensitive approaches is likely to be an essential component to reaching UNAIDS fast-track targets for 2020.
HIV self-testing (HIVST) appeals to the very people left behind by existing HIV testing services (HTS), including young people (15 to 24 years), adult men, key populations (men who have sex with men, people who inject drugs, people in prisons and other closed settings, sex workers and transgender people) and partners of people living with HIV (PLHIV). HIVST provides an empowering opportunity for individuals to test when, where and with whom they want to [14]. The ability to test in private and having more control over the testing process have been cited as key motivators to self-test particularly among men and young people [13,14]. When followed by timely uptake of prevention and treatment services, HIVST can be a key element in the push towards ending AIDS [15,16]. While previous studies have reported on preferences and uptake of HIVST, there has yet to be a multi-country investigation into the impact of alternative distribution and linkage strategies to optimize testing, VMMC and treatment after HIVST among men. Here, we present quantitative programme data from different HIVST distribution models.

| METHODS
Distribution models are summarized in Table 1, with five main approaches described below. OraQuick HIV Self-Test (OraSure Technologies LLC, Bethlehem, PA, USA) kits were distributed in all countries. Data reported here relate to the first 15 months of distribution (May 2016 to July 2017).
Social harms monitoring systems were part of all distribution models. No suicides were identified and reports of other serious harms (potential life-threatening/life-changing) were rare (1 event per 10,000 HIVST kits distributed), as discussed in detail for Malawi in this JAIS Special Issue [17].

| Model 1: community-based HIVST distribution
Community-based distributors (CBDs) provided HIVST kits across 53 districts in Malawi, Zambia and Zimbabwe. Models are described in detail elsewhere [18][19][20]. In brief, CBDs needed to have completed secondary school education and be resident in the distribution community. CDB recruitment used participatory approaches with candidates nominated following community sensitization meetings. CBDs completed a two-day training provided by Population Services International (PSI) including basic facts about HIV transmission and treatment, antibody-based diagnosis, discordancy and the principles of consent and confidentiality, as well as familiarization with the kits and how to demonstrate use to recipients, and data capture tools. All trainees had to undergo competency testing at the end of the training course when training skills were assessed. CBDs promoted and offered free HIVST kits for use alone or with CBD support. The same methods were used by CBDs to offer HIVST kits in households and social venues such as market places, busy streets, bars and beer halls. Individuals could also collect kits from the CBDs home at any time, if preferred.
CBDs provided all clients with brief health information about HIV, information on the test, and an in-person or videoclip demonstration-of-use and instructional materials optimized for local use demonstration to supplement manufacturer's instructions-for-use that were available in local languages.
Clients could choose to self-test alone, or with the CBD, and were asked to return their used kit and results in a sealed envelope, together with a short, self-administered questionnaire (SAQ) in collection boxes at community locations. Illiterate and semi-literate participants were supported by the CBD who was reading out the questions and answers from the SAQ with participants then left to complete the check-box answers in private.
Additional post-test guidance was available from CBDs on demand. All self-testers received self-referral cards with several locally adapted options to facilitate results-based linkage into HIV care and prevention services. CBDs collected information on social harms related to HIVST and referred clients for additional management as needed. A toll-free hotline was available to answer questions about the testing process, results and referral options.

| Model 2: HIVST integration into PSI-led HTS facilities and mobile HTS outreach
Integrated HIVST was piloted from June 2016 and scaled-up from January 2017 as an alternative option to provider-delivered testing for clients attending existing PSI-led HTS clinics and 11 mobile outreach sites in Zimbabwe. Outreach sites included "hot spots" at bus and truck stops, mining areas and urban shopping malls, and other informal workplaces. The aim After registration, HTS clients were offered a kit that they could use for HIVST on-site or at home. Clients opting for HIVST received a brief demonstration either by video or by a trained provider. Clients opting out of HIVST received conventional HTS. Private cubicles or tents, with offer of counsellor assistance, were provided to those self-testing on site. On-site confirmative testing was available for those reporting a reactive (positive) self-test result. If confirmed, PLHIV were referred for ART according to national guidelines, with immediate initiation if ART services were either available onsite, or through a referral form to ART services at public and private sector health care facilities.
All clients opting for HIVST received information about post-test support services and referral forms (confirmative testing and HIV treatment including ART for those with reactive results, information about prevention services for those with negative HIVST results) prior to HIVST. Men were encouraged to consider VMMC if they tested negative, and condom use was promoted. Clients who decided to self-test at home received information materials listing local prevention and treatment services, and a self-referral form suitable for either prevention or ART services, dependant on HIVST result.
HIV positive index clients diagnosed at the HTS site were offered self-test kits for secondary distribution to all their sexual partners for the purposes of index-testing [21]. Clients taking kits for secondary distribution were talked through the process of supporting their partner to use and interpret the kit correctly, how to access follow-on HIV services, and the need to maintain voluntariness [22].
Self-testers were asked to leave their used test kits with an SAQ in sealed envelopes at the site, while provider-delivered HTS clients had data captured by the counsellor. Used selftest kits were re-read by the providers on the same day, with this approach used to estimate the number and proportion of HIV-positive self-tests.

| Model 3: HIVST distribution at workplaces
At larger male dominated workplaces in the mining and farming industry, HIVST kits were distributed through peer-promotors or PSI HTS outreach workers, who provided pre-test information and in-person demonstrations of the self-testing process. Clients could self-test on site or at home and could take a test kit home for their partner to use, with support for secondary distribution as described above. Confirmatory testing was available on site, provided by the PSI HTS outreach team or by workplace HTS services, or through self-referral forms providing information on local private and public-sector health services. Confirmed PLHIV were referred for ART at public or private sector providers. A toll-free hotline number was provided to all clients.

| Model 4: HIVST distribution at public sector health facilities
Patients accessing public sector outpatient departments (OPD) or other clinical services were offered HIVST by healthcare providers, either nurses or counsellors working at OPD, before their consultation. Clients could self-test in a separate room following a brief demonstration, with the option of sharing their results during their consultation. Information on confirmatory testing, ART and HIV prevention services was provided to all patients. For those with positive selftests, counselling, confirmatory testing and ART were available on-site through the routine facility services. HIVST-negative clients received HIV prevention messages by the nurse and healthcare provider in OPD and male clients were referred for VMMC.
2.5 | Model 5: HIVST integrated with VMMC promotion VMMC was already being rolled-out in all three countries by PSI, and HIVST was integrated into mobilization strategies. VMMC mobilizers were trained to offer HIVST to all men who were interested in circumcision, but cited fear of HIV testing onsite. VMMC mobilizers, who had all received a two-day training course, as described for the CBDs, provided pre-test information and demonstration of kit use before offering a kit to each potential VMMC client. In Zambia, VMMC mobilizers also distributed HIVST kits to women.

| Data collection and analysis
HIVST kit distributors collected individual-level demographic and HIV testing history data from all clients, using either electronic or paper-based forms. Data from SAQs were entered into databases at country-level. Data were aggregated and presented by distribution model at PSI central level. STAR HIVST programme data from Malawi, Zambia and Zimbabwe was analysed according to age, sex, distribution model, testing history and compared between countries. We also compared characteristics of clients, including HIV result and number of HIV-positives identified, who took up the offer of HIVST with those of clients preferring provider-delivered HTS at PSI-led facilities and mobile outreach services. Given the high numbers of testing events (making standard p-values uninformative), and the intrinsic clustering nature of data from different sites, we present data descriptively without use of testing for statistical significance.

| Ethical considerations
All HIVST kits distributed before July 2017 were covered by country-level research protocols approved by the Ethics Committees of London School of Hygiene and Tropical Medicine, and the relevant ethics committees in Malawi, Zambia and Zimbabwe. As a public health intervention using a version of an HIVST product already approved for over-the-counter sale in USA and shown to have minimal potential for harm in Malawi, approved protocols included request for waiver of written or verbal informed consent for HIVST clients. Clients were instead informed about the investigational nature of the HIVST kit through community sensitization events, information leaflets and marking of kits as for research purposes only. Other models were delayed by need for initial piloting, and some (notably HTS integration and VMMC demand creation) were also dependent on the scope and scale of suitable PSI programmes, which varied country-to-country. In this respect, Zimbabwe-PSI had a large HIV service provision platform from which to rapidly diversify and scale-up HIVST models based on integration into fixed and outreach teams already providing HTS, accounting for 52,254 of the 54,453 kits distributed using this model to July 2017. Similarly, in Zambia, the large pre-existing VMMC programmes supported rapid scale-up of HIV delivered through VMMC mobilizers (15,092), with Zambia also leading on integration of self-testing into public sector clinics.

| Reach to first-time testers
The overall proportion of first-time testers (Tables 2 and 3) was 19.6% (119,673), varying from 26.8% in Malawi, to 21.6% in Zambia, to 13.6% in Zimbabwe (where self-testing was introduced to communities previously served by standard HTS delivered by mobile outreach teams). A higher proportion of men (overall 22.3%) than women (overall 17.1%) were firsttime testers in each of the three countries.
A further breakdown of the proportion of all self-testers who were first-time testers is shown for men and women by age-group in Table 3. This shows higher proportions of firsttime testers in the youngest age-group for both young men (29.4%) and women (24.4%), but with a substantial minority of clients in the older age-groups for both men (16.4% to 17.1%) and women (10.6% to 15.1%).

| Community-based distribution model
The CBD model was evaluated in detail for safety and population-level impact, with social harms monitoring and household surveys conducted to evaluate coverage and linkage, as reported elsewhere [17][18][19][20]. Use of distributed kits was confirmed for 275,419 (53.0%) by return of used kits, with country-level data for this variable being 53.2% (86,925) in Malawi, 58.8% in Zambia (92,247) and 48.2% in Zimbabwe (96,247).
CBD models varied substantially country-by-country [21][22][23], with the Zimbabwe model being based on delivery from mobile teams that supported training and brief (three to four weeks) but intensive HIVST distribution by temporarily employed distributors. CBDs in Malawi and Zambia were employed for 12 months to provide services less intensively. Recruitment and training are summarized under methods. The number and age of recruited distributors are shown in Among those self-testing, 1908 (859 men and 1072 women) were newly diagnosed with HIV. Provider HTS clients had a substantially higher HIV prevalence (10.2% positive) than self-testers (1.9% positive).

| Other models of distribution
Other models (Table 2) included public sector facilities in Zambia (45.8% men) and Zimbabwe (29.0% men) and workplace distribution (9850 kits) in Malawi and Zimbabwe, with over 66.4% and 58.9% of HIVST kits taken by men.
A total of 23,561 tests were distributed to men reached with mobilization for VMMC in Malawi (1327), Zambia (15,092) and Zimbabwe (7142). Referral tracking data from Zimbabwe showed that 40.2% of males who had received HIVST kits prior to VMMC went on to be circumcised.

| DISCUSSION
STAR is the largest evaluation of HIVST implementation to date. With 628,705 kits distributed in Malawi, Zambia and Zimbabwe within 15 months of introducing HIVST as a novel approach at community and facility-level, acceptability was high. We used five main distribution models, although community-based distribution accounted for 82.7% of kits distributed. Approximately half of all HIVST participants were men, with good male representation in all distribution models and age groups. A substantial minority of participants had never tested for HIV before, with this proportion higher for men (22.3%) than women (17.1%), and higher for young people (16 to 24 years: 26.9% first-time testers) than older age-groups.
HIVST is a promising approach for reaching underserved subpopulations who have never tested before and contributing to the realization of the UNAIDS fast-track strategy.
Consistent with previous reports [15,16,22,23], all distribution models had high male participation in each country. Strategies that provide men with greater coverage of HIV testing and care are urgently needed both to address the disproportionately high testing gap and mortality from HIV in men, and also to reduce risk of onward transmission of HIV [3][4][5][6][7]25]. Peak HIV prevalence for men in southern Africa is now in the 40-to 49-year-old age-group [3][4][5][6][7], with older men among least likely to have accessed standard HIV testing services [5][6][7]. Older men appear relatively receptive to HIVST, however, as evidence by the data reported here as well as from implementation studies from Kenya, Lesotho and Zimbabwe [15,[26][27][28]. For adolescent boys, HIVST can provide the first opportunity to test without fear of judgement from parents and healthcare workers [14], explaining the high uptake among this age group when HIVST was offered at community level. Thirty five percent of adolescent boys accepting self-testing were first-time testers in the STAR project in Malawi [14].
The STAR CBD distribution model was evaluated using household surveys, with uptake providing a measure of acceptability. Community-level coverage of HIVST was 42.5% of all surveyed adults in rural Malawi [29], and 50.3% in rural Zimbabwe [19]. This type of community-based HIVST distribution could then contribute to activities such as national HIV testing campaigns, targeted "catch-up" campaigns in districts with low testing coverage, and as a way of providing ongoing or periodical HIV testing access in remote communities. Costs (range US$7.23 per kit distributed in Malawi, to US$14.58 in  Zambia) and affordability of the CBD model are discussed in the accompanying manuscript by Mangenah et al. [24], and the potential to devolve HIVST further through communityled approaches is currently under investigation within STAR. Community-led models can deliver better outcomes at or below the cost of less integrated approaches and are widely used in Africa for mass drug administration and distribution of insecticide-treated bed nets [30].
Integrating HIVST into routine HTS services and in clinical settings, where access barriers may preclude testing by everyone that requires it, also shows promise with over 80% of men and women accepting HIVST when offered as an alternative to provider-delivered HIV testing. Our data, alongside that from alternative models of integrated facility-based HIVST [31], suggest that HIVST can contribute substantially to comprehensive provider-initiated HTS in high volume and congested public sector clinics [31]. This model could be expanded to public sector healthcare facilities more widely, especially where the current testing capacity is limited or poorly implemented [31]. We also show marked preference for HIVST in all fixed and outreach HTS sites where HIVST was offered as alternative to standard HTS. Preference for HIVST was most pronounced when queuing was needed to access standard HTS, but was also apparent in home-based testing services where 61.9% of clients who tested opted for HIVST, supporting other suggestions that HIVST is generally preferred by many Africans [14,16].
Integrated HIVST offers potential efficiency gains, minimizing time commitments for clients and streamlining management for providers in part due to much lower HIV prevalence in those opting for HIVST over HTS. This suggest a pronounced "self-selection" as noted for sex workerswho mostly opted for standard HTS when entering dedicated sexual health services [32] before considering HIVST for subsequent tests, citing potential higher sensitivity of blood-based provider-delivered tests as being important to them given their high exposure. Although speculative, and needing further research to confirm this, if our data do indeed reflect selfselection with individuals at low risk for HIV more likely to opt for HIVST, then this has a number of advantages. From the perspective of service providers, this allows for task-sharing with low-risk clients, allowing counsellors to focus their time on the remaining clients with a high risk of being HIV positive, and to dedicate more time on more time-consuming testing options such as index testing and assisted partner notification.
Introducing the option of HIVST greatly increased the numbers of clients who could be served each day at rural and urban outreach services and consequently increased the number of positive cases identified per counsellor and per site at any given time [33]. A further low-cost facility-based model ("secondary distribution"), where HIVST kits can be delivered to partners by antenatal clinic attendees and newly diagnosed PLHIV [22,27] is being scaled-up under STAR in Malawi and Zimbabwe and is discussed further in the accompanying manuscript relating to social harms in this issue [17].
This study has a number of limitations. This analysis is based on programmatic data from three different countries and is based on self-reported client data, with some missing data. Data on HIVST with regards to first-time testers, motivators and barriers to HIVST may have been prone to social desirability bias. As reporting on first-time testing was based on a subset of self-test users who had returned their used tests together with the questionnaire, responses might not be representative for the entire HIVST population. For the difference in HIV prevalence in our integrated HTS model, we cannot exclude alternative explanations, including that some HTS clients were obtaining confirmation of an earlier positive test or self-test, as many clients coming in for HTS are reluctant to detail previous positive results for a variety of reasons. Finally, the results may not be generalizable to other programme contexts with less intensity of distribution or different starting attitudes and perceptions by potential HIVST users and HTS providers.   HIVST, HIV self-testing.

| CONCLUSIONS
Men and young people in sub-Saharan Africa contribute disproportionately to the number of PLHIV who are not aware of their status. Results from two years of large-scale implementation of HIVST through several distribution models demonstrate how targeted roll-out could increase coverage of HIV testing, contribute to case finding among difficult to reach priority populations, particularly among high-risk men and young people and increase efficiency and capacity of HTS in high volume and overcrowded clinics. HIVST offers clear advantages when provided in addition to existing services, and if scaled-up, can contribute to closing the gap towards the "first 90. "