Uptake of Home-Based HIV Testing, Linkage to Care, and Community Attitudes about ART in Rural KwaZulu-Natal, South Africa: Descriptive Results from the First Phase of the ANRS 12249 TasP Cluster-Randomised Trial

Background The 2015 WHO recommendation of antiretroviral therapy (ART) for all immediately following HIV diagnosis is partially based on the anticipated impact on HIV incidence in the surrounding population. We investigated this approach in a cluster-randomised trial in a high HIV prevalence setting in rural KwaZulu-Natal. We present findings from the first phase of the trial and report on uptake of home-based HIV testing, linkage to care, uptake of ART, and community attitudes about ART. Methods and Findings Between 9 March 2012 and 22 May 2014, five clusters in the intervention arm (immediate ART offered to all HIV-positive adults) and five clusters in the control arm (ART offered according to national guidelines, i.e., CD4 count ≤ 350 cells/μl) contributed to the first phase of the trial. Households were visited every 6 mo. Following informed consent and administration of a study questionnaire, each resident adult (≥16 y) was asked for a finger-prick blood sample, which was used to estimate HIV prevalence, and offered a rapid HIV test using a serial HIV testing algorithm. All HIV-positive adults were referred to the trial clinic in their cluster. Those not linked to care 3 mo after identification were contacted by a linkage-to-care team. Study procedures were not blinded. In all, 12,894 adults were registered as eligible for participation (5,790 in intervention arm; 7,104 in control arm), of whom 9,927 (77.0%) were contacted at least once during household visits. HIV status was ever ascertained for a total of 8,233/9,927 (82.9%), including 2,569 ascertained as HIV-positive (942 tested HIV-positive and 1,627 reported a known HIV-positive status). Of the 1,177 HIV-positive individuals not previously in care and followed for at least 6 mo in the trial, 559 (47.5%) visited their cluster trial clinic within 6 mo. In the intervention arm, 89% (194/218) initiated ART within 3 mo of their first clinic visit. In the control arm, 42.3% (83/196) had a CD4 count ≤ 350 cells/μl at first visit, of whom 92.8% initiated ART within 3 mo. Regarding attitudes about ART, 93% (8,802/9,460) of participants agreed with the statement that they would want to start ART as soon as possible if HIV-positive. Estimated baseline HIV prevalence was 30.5% (2,028/6,656) (95% CI 25.0%, 37.0%). HIV prevalence, uptake of home-based HIV testing, linkage to care within 6 mo, and initiation of ART within 3 mo in those with CD4 count ≤ 350 cells/μl did not differ significantly between the intervention and control clusters. Selection bias related to noncontact could not be entirely excluded. Conclusions Home-based HIV testing was well received in this rural population, although men were less easily contactable at home; immediate ART was acceptable, with good viral suppression and retention. However, only about half of HIV-positive people accessed care within 6 mo of being identified, with nearly two-thirds accessing care by 12 mo. The observed delay in linkage to care would limit the individual and public health ART benefits of universal testing and treatment in this population. Trial registration ClinicalTrials.gov NCT01509508

will be performed during 2013 to inform decision-making regarding process and procedures for phase 2.
Sample size: A fully parameterised, deterministic mathematical model demonstrates that a 34% reduction in cumulative HIV incidence (4.5% versus 2.98%) in HIV-negative participants over two years should be feasible across a wide range of parameter space. Sample size calculations indicate that 22 clusters (11 in each arm), with 1 000 consenting participants >15 years of age in each cluster (N=22 000; 17 600 HIV-negative), are required to achieve this objective. The first phase will be conducted on 10 000 participants in 10 clusters, which allows the measurement of the proportion agreeing to test over three rounds of testing within 1% (95% CI) and uptake of testing in the intervention communities of all HIV-positive participants within 4% (95% CI).
Expected results: We aim to provide proof-of-principle evidence regarding the effectiveness of Treatment-as-Prevention in reducing HIV incidence at the population level. We will collect data from the participants to inform the generalizability of the results, and thus inform policy resulting in wide implementation.
Objectifs : Estimer directement l'impact du traitement ARV initié immédiatement après le diagnostic de l'infection par le VIH et quel que soit le niveau de CD4 des patients non encore éligibles au traitement ARV, sur l'incidence de nouvelles infections VIH dans la population générale de la même région.

HAART
Highly Active AntiRetroviral Therapy     (3); the recent results from the Tenofovir (TDF) -containing microbicide are promising but need confirmation in a further trial (4). Treatment of sexually transmitted infections, including herpes simplex has no measurable effect on HIV transmission (5). Male circumcision halves the risk of HIV transmission from females to males but is not (yet) a widely-used public health measure, although the South African government has started implementing a programme, focussing on young men in April 2010. As a consequence, effective prevention continues to rely on behavioural change and condom use, but these methods have their limitations, as evidenced by the resurgence of unsafe sex in homosexual males in Western Europe and the USA (6), and limited uptake and more importantly sustainability in lower-income countries (7,8). At their current or anticipated level of use none of these methods is effective enough to contain the pandemic in countries with high incidence and prevalence of HIV infection such as South Africa.

Antiretroviral therapy can reduce sexual transmission
HIV plasma viral load (VL) in the index HIV-infected individual is the dominant determinant of transmission, documented in heterosexual couples and mother/child pairs (9,10). Antiretroviral therapy (ART) with fully suppressive ARV drugs combinations lowers VL in all body compartments and decreases the risk of transmission to a very low level. It is thus legitimate to raise the following question: Could ART contribute to reducing transmission at individual and population levels? Head-to-head comparisons are not available and are unlikely to be promoted, but in cohort studies where routine use of condoms was encouraged in sero-discordant couples, results differed substantially depending on whether index patients were on ART (0.46 case of HIV acquisition per 100 person-years [ppy]) or not (5.64 ppy) (11). In the same meta-analysis, Attia et al estimated a zero (97.5% upper confidence limit of 1.27 ppy) transmission risk for those on ART and with successful viral suppression (VL <400 copies/ml). The HPTN 052 study showed that ART reduces HIV transmission by 96% in stable couples where one partner is HIV-infected and the other is not (12). This NIH trial has not addressed the question of the reduction of HIV transmission at the population level.

Can ART be provided at earlier stages of HIV infection than currently recommended?
Until 2009 international recommendations for initiating ART were generally conservative everywhere in the world, especially in resource-limited settings where ART initiation was recommended only at a CD4+ (CD4) count of 200 cells/mm 3 or less or for end-stage HIV disease. However, recently two pieces of evidence considerably changed this approach: analyses of observational data from the USA and Europe (13,14) and a randomized clinical trial conducted in Haiti (15). Further, there is some evidence of benefit of earlier treatment from trials of structured treatment interruption. For instance, the ANRS 1269 TRIVACAN trial conducted in Côte d'Ivoire showed that patients interrupting ART at intermediate CD4 levels had higher risks of morbidity and mortality than patients remaining on ART (16). Finally, observational cohort data from Zimbabwe have shown that the risk of death of untreated HIV-infected women was 6.2 times higher for those with more than 600 CD4 when compared with HIV-negative women (17).
In November 2009, WHO recommended to substantially broaden eligibility for ART, with treatment of all HIV infected people with CD4 <350 cells/mm 3 irrespective of clinical symptoms or at WHO clinical stage 3 or 4 irrespective of CD4 count. Treatment should also be provided for those with a diagnosis of active tuberculosis (TB) irrespective of CD4 cell count, those who have a coinfection with hepatitis B virus (HBV) if the latter requires treatment and for pregnant women fulfilling treatment criteria (18). Two clinical trials are currently investigating whether ART initiated well above 350 cells/mm 3 provides sufficient additional individual benefits in terms of mortality and severe morbidity: the NIH-sponsored START trial (clinicaltrials.gov identifier NCT00867048) and the ANRS-sponsored TEMPRANO trial (clinicaltrials.gov identifier NCT00495651). Although the results of these trials are not expected before the end of 2015, evidence and programmatic experience have continued to shift the risk-benefit ratio towards starting ART earlier. Thus in June 2013, the revised WHO guidelines promoted "expanded eligibility for ART with a CD4 threshold for treatment initiation of 500 cells/mm 3 or less for adults, adolescents and older children. Priority should be given to individuals with severe or advanced HIV disease and those with CD4 count of 350 cells/mm 3 or less. ART is recommended to be initiated regardless of CD4 count for certain populations, including people with active tuberculosis (TB) disease who are living with HIV, people with both HIV and hepatitis B virus (HBV) infection with severe chronic liver disease, HIVpositive partners in serodiscordant couples, pregnant and breastfeeding women and children younger than five years of age" (19).

Strengthening HIV testing approaches
Without HIV testing, bio-medical interventions cannot target HIV-infected individuals. Motivations for testing vary, but many people test to obtain treatment, often when they feel unwell. Restricting treatment to those who have already developed symptoms and meet CD4 count criteria as discussed in the previous section may actually serve as a serious, if unintentional, barrier to the uptake of HIV testing.
Over the past two decades, HIV counselling and testing services have primarily been promoted and provided in the context of Voluntary HIV Counselling and Testing (VCT). However, the limits of VCT approaches in resource-limited settings, particularly regarding achieving early diagnosis, high levels of population coverage and access to "hard-to-reach" groups (20,21), have led to the development of other innovative forms of HIV testing such as Provider-Initiated Testing and Counselling (HIV testing and counselling recommended by health care providers for any persons attending health care facilities as a standard component of medical care, such as HIV testing in TB clinics, door-to-door testing, partner notification), home-based or mobile HIV testing that take testing services to people (22).
Defining a feasible, acceptable, and efficient strategy to obtain very high rates of testing (e.g. near universal HIV testing repeated bi-annually) may necessitate further development of both existing VCT services available and extension of PITC. The South African government strongly supports a coherent, consistent HIV services' approach, encouraging all public health facilities -fixed and mobile -to offer HIV testing. Successfully testing the numbers needed to achieve the coverage required for implementation of treatment for all or nearly all infected people is quite different to simply promoting VCT, and will certainly necessitate a more intensive approach to provider-initiated services, taking into account the needs and attitudes of different population groups and considering the relevance of different modes of HIV testing in different populations.

Could combination antiretroviral therapy be used universally to reduce sexual transmission of HIV at population level? If yes, which combination?
In 2009 a mathematical modelling exercise using a hypothetical population and assumptions relating to the South African setting concluded that "Universal voluntary HIV testing and immediate ART (regardless of CD4 count) combined with present prevention approaches, could have a major effect on severe generalised HIV/AIDS epidemics" (23). Two accompanying commentaries addressed how and whether this was feasible to deliver in the 'real world' (24,25). The modelling exercise showed that HIV transmission could be substantially reduced within a few years. The authors argued that current clinical practice, which in nearly every setting, relies on CD4 counts and advanced HIV disease as the trigger to introduce ART, limits its preventive efficacy by leaving key points in HIV's natural history to go unchecked by effective VL-reducing treatment (these key points are the peak VL at seroconversion and the sustained period of somewhat elevated VL set point during the asymptomatic period). Dodd et al developed a deterministic mathematical model to investigate the impact of test-and-treat interventions under a variety of assumptions about the epidemic (26). They showed that such an intervention could substantially reduce HIV transmission, but that the impact might depend on the epidemiological context (notably determined by the sexual partner network, such as heterogeneity, concurrency and mixing).
Providing that a universal test and treat approach is appropriate, the question of the choice of ARV drug combination for wider and prolonged use becomes central. This ARV regimen should fulfil the following criteria: 1) Appropriateness for all CD4-cell strata to simplify its use; 2) Minimal side effects in otherwise "healthy" patients to avoid treatment drop-outs; 3) High potency to maximize effect on transmission; 4) High genetic barrier to minimize acquisition of viral resistance; 5) Sustainability for many years to limit and delay switches to secondline ARV regimens; 6) Low pill burden to facilitate treatment adherence; 7) Minimal laboratory requirements for follow-up; 8) Safety; 9) Affordability and 10) Coverage of special populations (TB co-infection, hepatitis B co-infection, pregnant women). A fixed-dose-combination (FDC) of (efavirenz [EFZ 600mg / tenofovir disoproxil fumarate [TDF] 300mg / emtricitabine [FTC] 300mg) available in South Africa as Atripla® or other generic forms administered once daily fulfils most of these criteria.
Not only may earlier treatment reduce HIV incidence at a population level, it may also benefit the individual. The long-term benefits of starting ART earlier would likely be even more important in settings where the incidence of life-threatening HIV-related diseases occurring at relatively high CD4 levels (tuberculosis, invasive bacterial diseases, and possibly malaria) is substantial, a typical situation in most sub-Saharan Africa including South Africa. Two randomised clinical trials have been designed or are in progress to address the "when to start" question in different settings and at different CD4 levels. The TasP trial addresses the same issue but expands it further, although prevention of new infections is its "raison d'être" (26).

Study hypothesis
HIV testing of all members of a community, followed by immediate ART initiation of all HIV-infected individuals regardless of immunological or clinical staging will prevent onward sexual transmission and reduce HIV incidence in this population.

Study design
TasP will be a population-based cluster-randomised trial (RCT), with small communities used as the units for randomisation. It aims at evaluating whether ART delivered at an earlier stage of HIV infection than currently recommended can prevent onward transmission of HIV. This effect on HIV incidence among HIV-negative subjects will be the direct consequence of the treatment initiation among those non-eligible for ART according to the current standards. TasP aims to address this very important question and produce such evidence, from a rural area in KwaZulu-Natal, the province with the highest HIV incidence and antenatal prevalence in South Africa, the country with one of the highest HIV burdens in sub-Saharan Africa, the continent where 70% of new HIV infections occur.

Trial phases
The TasP trial will be implemented in a two-phase approach:  The first phase will be conducted in a selected number of clusters (see section 8). All trial procedures to be implemented in this first phase will be as planned in this full trial protocol.  If in the first phase the procedures and approach are shown to be feasible and acceptable and if the aims of the trial are still deemed relevant within the context of international research advances and the research strategy of the Africa Centre and partner institutions, and in agreement with the Steering Committee (SC), recommendation of the Scientific Advisory Board (SAB) and the Data Safety Monitoring Board (DSMB), the protocol will be implemented in the remaining clusters (see section 8) in the second phase.

Expected results of the trial
We aim to contribute to the evidence-base that a Treatment as Prevention approach can be applied to an entire population with a high incidence of HIV infection, irrespective of the stage of HIV disease and circumstances of HIV testing of participants in that population, and that Treatment as Prevention contributes to a reduction in HIV incidence while yielding further population and individual benefits.
The evidence obtained should be strong enough for this approach to be replicated in a wide range of resource-limited settings, to have policy implications to result in wide implementation.

Hlabisa sub-district
The trial will be conducted in Hlabisa sub-district, Umkhanyakude district, Northern KwaZulu-Natal, South Africa ( Figure 1). The Hlabisa health sub-district is a rural setting of 1 430 km 2 in size, with a population of approximately 220 000 Zulu-speaking people of whom 3.3% are located in a formal urban township (KwaMsane), 19.9% in peri-urban areas and the remainder (76.8%) classified as living in a rural area. The rural population lives in scattered homesteads that are not concentrated into villages or compounds (as would be the case in many other parts of Africa). Health facilities in the Hlabisa sub-district are provided by one central community hospital (300 beds) and 17 fixed primary health care clinics (Department of Health clinics), which provide the bulk of the health care for the population of the sub-district. All clinics provide care for minor ailments, family planning services, antenatal and postnatal care, conduct deliveries, treat sexually transmitted diseases, provide child immunisations, diagnose and manage TB, and care for chronic illnesses such as diabetes and hypertension (and HIV treatment and care, see below section 2.3). In addition to the 17 fixed clinics, 31 mobile clinic points are provided twice monthly, mainly for childhood immunisations and maternal and child health care. The sub-district is also serviced by approximately 130 community health workers, each of whom is expected to regularly visit a group of assigned homesteads.

The South Africa HIV treatment programme
The HIV treatment programme in South Africa is the largest in the world, with 2 million individuals estimated to have been receiving HAART in mid-2012 (27) (HSRC July 2013). One key aim of the current National Strategic Plan for HIV is to provide access to treatment, care and support to 80% of HIV-infected people and there has been significant scale-up of activity in the past few years (28). There remains, however, a large unmet need for treatment, with an estimated coverage in 2009/2010 of approximately 58% of those eligible for treatment, although this varies across districts (28).
The national response to HIV and the development of implementation plans is informed by the national strategic plan on HIV, STIs and TB (28) (ref) and the delivery of ART is guided by the recently updated National Antiretroviral Treatment (29). Until early 2010, HIV-infected adults were eligible for ART if CD4 cell count <200 cells/mm 3 or WHO stage IV condition. In April 2010, new guidelines were implemented where pregnant females and individuals with TB disease became eligible for ART if CD4 cell count <350 cells/mm 3 , and in August 2011 these guidelines were further extended to eligibility starting at 350 for all HIV infected people in addition to all those with WHO stage 3 or 4 disease or anyone with MDR or XDR TB. The 2013 South African revised antiretroviral treatment guidelines also include all pregnant women to be eligible for ART for the duration of pregnancy and breastfeeding (29) (see Appendix 15.1). The current first-line drug regimens in use in South Africa are TDF + 3TC/FTC + EFV/NVP (see Appendix 15.1).

Hlabisa HIV Treatment and Care Programme
The Hlabisa HIV Treatment and Care Programme is a partnership between the KwaZulu-Natal Department of Health (DoH) and the Africa Centre, aiming to provide an accessible, equitable and comprehensive service to all people living with HIV infection (30). Established in 2004, the service is decentralized to all 17 primary health care clinics in the sub-district, and is nurse-and counsellordriven, with scheduled physician visits to initiate patients on treatment and manage clinical problems. Recently, agreement was reached regarding nurse-led initiation and a number of nurses within the programme are enrolled on training programmes to allow this to happen. By September 2011, over 18,000 HIVinfected people eligible for treatment had started HAART; patients' eligibility is determined by South African guidelines (see section 2.2). There are no waiting lists for treatment initiation, approximately 250 adults and 30 children are started on treatment each month, and around 2,500 voluntary counselling and testing sessions are conducted monthly in the clinics or within the mobile and homebased testing services. In addition to those on treatment, more than 20,000 individuals are being monitored prior to requiring therapy. The Hlabisa HIV Treatment and Care Programme also includes TB screening and treatment, PMTCT, PITC and VCT services, the latter delivered in the fixed clinics, as well as through mobile and home-based units (30). For individuals on ART between 2004 and 2008, retention in care at one year in the programme was 84.0% and mortality in the first year was 10.9% (31). Recent evidence shows a 46% decrease in early mortality for those who initiated ART in 2011/12 compared to the reference period 2008/9 (32).

The Africa Centre for Health and Population Studies
Hlabisa sub-district hosts the Wellcome Trust-funded Africa Centre for Health and Population Studies, a research institute of the University of KwaZulu-Natal, with a focus on HIV epidemiology and prevention (www.africacentre.com); the Centre partners with the Department of Health in the delivery of HIV treatment and care in the Hlabisa sub-district. A core activity of the Africa Centre is the longitudinal socio-demographic (bi-annual) and HIV (annual) surveillance in a geographically defined area in the south of the sub-district, covering about 40% of the area and of the population in the sub-district -the Demographic Surveillance Area (DSA) (see shaded area of Figure 1).
The TasP trial will take place in the sub-district outside of the Africa Centre surveillance area, so that the on-going HIV surveillance in the Africa Centre area can continue to monitor changes in behaviours and risks over time in a population very similar to the one where the trial will take place. The trial area contains six of the 17 fixed primary health care clinics in the sub-district.

Overall objectives of the trial (first and second phase)
To compare the effect of ART initiated immediately after HIV diagnosis, irrespective of CD4 count criteria versus South African guidelines, on the reduction in incidence of new HIV infections in the general population in the same setting.

Specific objectives
Among all participants  To compare the acceptability and feasibility of providing HIV testing to all members of a community between the two trial arms, and more specifically:

Main objective 
To validate and update the parameters of the model used to estimate the trial sample size and HIV incidence, in terms of: uptake of HIV testing, linkage to care upon HIV diagnosis, internal migration and ART initiation.

Specific objectives
Among all participants  To estimate the acceptability and feasibility of providing HIV testing to all members of a community, and more specifically:

First phase
The first phase of the trial is planned over a 24-month period (see section 4.2). The aim of this first phase is to validate the hypothesis defined for the overall trial design (number of clusters, number of participants, incidence, HIV prevalence) and to verify the feasibility and acceptability of the intervention within the community.
The first phase will be conducted on a limited number of participants (n=10 000) and a limited number of clusters (2×5, see section 8).
In the initial four clusters, three rounds of home-based HIV testing of six, four and four months will be conducted. HIV-infected participants identified in all clusters will be followed-up between 7-24 months depending on the time of entry in the trial (first or third HIV testing rounds, see Figure 4).
For the additional 6 clusters, 2 rounds of home-based HIV testing of six months each will be conducted and this will coincide with the last 12 months of follow-up in the initial 4 clusters. Follow up of HIV-infected participants in these 6 clusters will be 1-12 months in the first phase.

Second phase
If in the first phase the procedures and approach are shown to be feasible and acceptable, if the aims of the trial are still deemed relevant within the context of international research advances and the research strategy of the Africa Centre and partner institutions, and in agreement with the Steering Committee (SC) and recommendation of the Scientific Advisory Board (SAB) and the Data Safety Monitoring Board (DSMB), the trial will be rolled-out to the remaining clusters (n=12, see section 8) in the second phase.
Four HIV testing rounds of six months each will then be conducted in all 22 clusters. HIV infected people from the phase 1 will continue to be followed in the TasP trial clinics until the trial ends when their care will be transferred to the programme as appropriate.

Description of the trial participants
Trial participation will be offered to all individuals meeting the following criteria:  aged ≥16 years;  member of a household in the designated cluster (head of household defines the membership status in Zulu culture);  able and willing to give written informed consent for trial participation and/or HIV counselling and testing.
Note: Individuals considered unable to provide informed consent will include those with severe uncontrolled psychiatric disorders, and those with neurological impairment resulting in an inability to participate in the informed consent process.

Trial components
The TasP trial consists in HIV testing of all members of the community at regular intervals (component 1) and comparing two ART initiation strategies for HIVinfected participants (component 2) as summarised in Figure 3 below.

Component 1: HIV counselling and testing and comprehensive prevention programme
This first trial component will be identical in both intervention and control clusters:  Provision of HIV counselling and testing to all members of the trial clusters (see section 6.2.2)  Access to a full set of preventive services for all trial participants, services already available in DoH clinics for HIV-negative participants and made available within the trial clinics for HIV-infected participants; this will include Information, Education and Communication (IEC) and condom distribution, circumcision services, syndromic management of STIs and post-exposure prophylaxis, family planning (see section 6.3)

Component 2: ART for HIV infected participants
Within the second component, HIV-infected patients identified within the trial will be offered ART according to different eligibility criteria between the intervention and control clusters:  in the intervention clusters: all HIV-infected adults will be offered ART regardless of their immunological and clinical staging;  in the control clusters, HIV-infected people will be assessed clinically and immunologically and when eligible for treatment as per South African guidelines will be offered ART.

Preparing the community
At the beginning of phase 1, a detailed community entry, education and trial promotion strategy that builds on the Africa Centre's extensive community engagement programme will be implemented based on the preparatory fieldwork conducted during late 2010, funded by the ANRS and the Africa Centre.
Key elements of the Africa Centre's existing and ongoing community engagement programme that provide opportunities to promote the trial include: bi-weekly sponsored community 'edu-tainment' road shows; annual community soccer and netball tournaments; annual community 'fun-run'; a widely circulated community magazine in isiZulu (Umbiko); participation in a local community radio and a new programme of 'edu-tainment' music CDs distributed and played by operators of the local mini-bus taxi association.
The funded exploratory work involved a community consultation process using in-depth interviews and consumer research panels with a range of key community informants meeting several times during the pre-pilot phase to determine what additional community education and awareness activities are needed for the first phase. The resulting community entry strategy for the trial itself will be developed jointly by the investigators and the Africa Centre's Community Liaison Office and tested with the community research panels in the preparatory period of the trial (see section 7.4.5).

Approaching households, obtaining individual consent, and administrating population-based questionnaires
The procedures presented below apply to each round of HIV testing. They are the same for the intervention and control clusters, for phase 1 and for phase 2.
All households within each of the trial clusters will have been identified and GPS coordinates noted. A team of trial community testers, all of whom will be DoHtrained VCT counsellors, will approach the household and seek permission to enter from the household head (or most senior household member present if head is absent). After entering, the testing team will explain the trial and the procedures for testing and seek the permission of the household head to offer trial participation and HIV counselling and testing to adult members of the household. Specific information sheets for each component will be provided to all individuals (See Appendix 15.4).
Once the head of household has agreed to allow the team to enter the household, he/she will be asked to complete the Tasp household registration questionnaire (THR) on the netbook and the TasP household information assets questionnaire (HHI) (see section 7.4.1).
Once permission to proceed has been granted by the head of household, a private space will be identified and all individual adult household members will be invited to give written permission to complete the TasP  The content of IQ is described section 7.4.1.
Participants who give consent to the anonymous collection of blood for HIVrelated tests will have blood collected by field workers by finger prick and stored on filter paper as dried blood spots (DBS) (see testing procedures section 6.4.6).
The HIV counselling and testing procedures are described below (see section 6.2.2).
In each round of testing, the team will be able to return on up to two more occasions to offer trial participation and/or HIV testing to members of the household who were not present during the initial visit, after which the untested members of the household will receive a written invitation to attend the trial clinics where HIV counselling and testing will also be available.
For participants who decline both HIV counselling and testing and the questionnaires, we will obtain basic demographic data from the household head. Participants who decline either testing or the questionnaire in the follow-up rounds will be asked to provide key, basic socio-demographic, knowledge of HIV status and HIV testing history data to the testing team (see section 7.4.1).

HIV counselling and testing procedures
Different HIV testing strategies currently operate in the trial area. HIV testing is available in DoH hospital and the 17 fixed primary health care clinics. This includes VCT clinics and PICT (provider initiated counselling and testing), including the routine testing of pregnant women and TB patients in clinics. In addition, some non-government organisations (NGOs) provide testing at fixed sites in Mtubatuba town and KwaMsane and may take testing to community venues and events. Further, the DoH/Africa Centre partnership programme, the Hlabisa HIV Treatment and Care Programme (see section 2.3), provides the largest and most comprehensive range of HIV testing services available, employing DoH-trained counsellors. The testing options available in the Hlabisa HIV Treatment and Care Programme include mobile clinics, offering testing in communities routinely, mobile testing clinics at community events, and homebased VCT offer in households across the whole of the Hlabisa sub-district, an initiative in the Programme started in 2009 which has proved to be highly acceptable with a high uptake rate (see section 1.1.4).
Within the TasP trial we will extend the provision of home-based VCT in households on a more regular basis with six monthly visits and combine the current range of community and clinic testing options, thus achieving a maximum (near universal) HIV testing coverage in the area and increasing options for repeat HIV testing.
During each HIV testing round, individuals providing written informed consent will receive pre-test HIV counselling privately and confidentially, delivered by a DoH trained counsellor.
Rapid HIV testing will be performed. The screening test will be Determine, and the confirmatory test will be First Response. Test results will be provided approximately 20 min after testing.
All participants who test HIV-positive will be referred to the trial clinic for further assessment, including a point-of-care CD4 count. Results from the baseline viral load test will confirm HIV positive status. Where there is discrepancy in results between the VL test and the rapid antibody test, the DBS from the home-based testing stored in the Africa Centre virology laboratory can be accessed for confirmatory antibody testing.
The individual post-test HIV counselling session will take place as per routine DoH procedures, covering the prevention of acquisition of HIV for negative people and the implications of HIV infection for positive people.
The same home-based HIV VCT procedures will be used in successive rounds of community testing to ensure a high uptake of testing and repeat testing overall, particularly as those who decline testing in one round become familiar with the procedures.
A standardised referral procedure will be developed to provide an opportunity to children and minors to be tested according to the South African guidelines at the DoH clinics and other facilities. Figure 4 summarises the home-based procedures at each round of testing.

Figure 4
Home visit chart.

Referral and linkage to care procedures
During each home-based VCT round, at the end of the counselling and testing session, all participants will be given the same referral/prevention card (regardless of their test result and/or parts of the trial they may have consented to). The referral card will contain information about the TasP trial clinics and the available services (treatment and care, VCT, condoms, referral to circumcision,

Head of household oral consent to enter
Household home-based questionnaire Referral to trial clinic etc.). Each participant will be asked to take their referral card with them when they attend clinics in order to identify them as a trial participant. A thumbprint will be taken at the TasP clinic to confirm a participant's correct identity and their correct service eligibility (i.e. treatment, care and prevention or VCT and prevention).
During post-test HIV counselling, HIV-infected participants will all be referred to the trial clinic in their cluster for further assessment as to their treatment eligibility (see section 6.4 below). HIV-infected participants will be informed that, if they have not attended the trial clinic within 3 months of referral, they will be contacted by a member of the study team (tracking team) either by phone or during a home visit for additional counselling and support to facilitate linkage to care.
A trained research psychologist will work with the tracking team to explore cognitive, psychological and normative aspects, which determine clinical linkage among individuals who test positive in the trial, but have not linked to care within 3 months of testing. All participants will be asked a screening question regarding the information received on the trial to test recall and understanding of opportunities to link to care. Thereafter psychological assessments will be completed using short validated screening tools with participants in their homes. This assessment includes psychological measures of depression, anxiety, somatisation (using the PHQ9; PHQ15 and GAD7) and health perceptions (MOS-36); normative aspects including stigma (HIV Stigma Scale -Revised) and family functioning (FAD). A brief cognitive screen (Oxford Cognitive Screen) on executive function will be administered using a tablet. This assessment will take approximately 45 minutes.

Preventive interventions
All participants, whether from intervention or control clusters, will have access to a full set of preventive interventions described below. HIV-negative participants will access these services, as does the general population, within DoH fixed clinics. All HIV-infected participants will be provided these services during their follow-up visits within the trial clinics.

Information and education, condom promotion and distribution
Specific behaviour change information and education, condom promotion and distribution activities, in line with what is currently provided in the Hlabisa HIV Treatment and Care Programme, will operate alongside those sponsored by national and provincial governments as well as various NGOs. VCT counsellors and clinic staff are trained in essential behaviour change counselling and available to provide this to trial participants both during VCT sessions or clinic visits. Social marketed condoms are currently routinely available through the DoH clinics as well as for purchase in small convenience shops (spaza shops) in the community, more formal shops, petrol stations and other outlets. On the basis of our audit and mapping of prevention and HIV testing services additional condom distribution plans for both intervention and control clusters will be implemented.

Male circumcision
The phased implementation of a national programme of male circumcision has started in South Africa and KwaZulu-Natal's roll-out started at test sites in 2010.
During 2011, the plan involved the setting up of mobile surgical facilities (in 'camps') where a trained physician performs the procedure for young adult males. The Africa Centre works closely with the DoH in implementing the programme in the sub-district as part of the Hlabisa HIV Treatment and Care Programme. However, since then the mobile camps have not been regular, and the programme has been taken over by NGO's working in the district; in addition circumcision is available from the district hospital, but there is limited capacity. Male circumcision is also available from private medical practitioners in the area and is covered by some employees' medical-aid schemes. Rates of male circumcision among participants in the Africa Centre's HIV surveillance have been low.

Syndromic management of sexually transmitted infections
All HIV-infected participants (male and female) will undergo a standardised STI symptom screen at the time of enrolment in the trial clinic, and at regular intervals once enrolled in HIV care. Symptom screen for women will include: lower abdominal pain, vaginal discharge, dysuria, and genital ulcers. Symptom screen for men will include: urethral discharge/dysuria, genital ulcers, and scrotal swelling/pain. This will be treated syndromically according to South African guidelines. The importance of partner treatment will be stressed and partner notification slips will be issued as per routine practice. All management will be in line with national guidelines and no additional diagnostic services will be included in the trial phase.

Post-exposure prophylaxis after sexual assault
All incidents of sexual assault for trial participants will be reported immediately to the police and management will follow standard guidelines. Post-exposure prophylaxis (PEP) will be offered to HIV-negative participants reporting penetrative anal or vaginal sexual assault who present within 72 hours. This consists of ZDV 300 mg bd + 3TC 150 mg bd (Combivir) and LPV/r (Kaletra) 400/100 mg bd for four weeks and will be administered either at Hlabisa hospital or at the existing crisis centre.

Family planning
All female participants will be asked about their use of family planning methods at the time of trial enrolment. Family planning advice will be offered to both HIVpositive and HIV-negative female participants, as part of the post-test counselling within the household. For HIV-infected people, there will be more opportunity for discussion at the trial clinics, whereas HIV-negative people will be referred to the DoH services. Dual methods (hormonal contraception + barrier) will be encouraged for all participants with no immediate pregnancy intentions. Any request to commence contraception will require referral to the fixed DoH primary health care clinic -this will be facilitated with a standardised referral form. HIV-negative participants will ordinarily be offered either oral contraceptives (several options) or injectable contraceptive (medroxyprogesterone acetate 150mg, 12-weekly). HIV-positive participants will be encouraged to use an injectable contraceptive.

Pre-treatment assessment visit
All participants newly diagnosed with HIV infection during any of the homebased VCT rounds will be referred to their trial clinic for immediate further assessment. Trial clinics will be informed about whom to expect.
Trial clinics (one per cluster) will be situated in close proximity to residences (< 45 minutes walking distance for all participants within the cluster).
These trial clinics will be staffed by a counsellor and a nurse. A referent physician will be available weekly at each clinic and on-call if necessary.
At the trial clinics, all newly documented HIV-infected participants will be provided standard counselling/ART education and adherence sessions by the ART counsellor and nurse, over one to three visits. They will be asked using separate documents to: 1) provide self-reported information and an anonymous blood specimen for viral load testing (see below) and 2) receive care and/or treatment at the trial clinics as per DoH standards. The HIV-infected participants consenting to treatment will then undergo a clinical evaluation which includes medical history, physical examination, WHO clinical staging (review of current and previous morbidity), basic anthropometry (weight and height), as per DoH procedures using the following CRFs: Clinic baseline visit -Counsellor (CBC), Clinic Follow-up -Counsellor (CFU), Clinic history and examination-nurse (CHE).
Patients will be referred to the ART counsellor to respond to the Social science clinic-based baseline questionnaire (SCB) and specifically questions regarding HIV testing experience, ART perception, disclosure and economic situation (see Table  3). This baseline interview will take place anytime during the pre-treatment visits.
In the intervention clusters, the clinical evaluation will involve completion of the following:  Blood tests: CD4 count (point of care*), VL, haematology (full blood count), biochemistry (urea, creatinine, electrolytes, liver function tests, glucose, lipids), hepatitis B surface antigen (HBsAg) and plasma storage,  Urine tests: dipstick urinalysis, beta hCG to all women of childbearing age.  Other screening: Sputum will be collected for TB investigation (microscopy and culture), as appropriate following screening on clinical symptom In the control clusters, the pre-treatment assessment will be decided after obtaining the CD4 count measured at point of care and/or according to the clinical staging (see ART initiation criteria in section 6.4.1 below):  Patients eligible for ART will undergo baseline investigations performed as above  Patients not eligible for ART will have a blood sample collected for baseline VL measurement and storage; they will be invited to return to the study clinic in 4-6 months based on the CD4 count All participants will be seen within two weeks of this W0 pre-assessment visit, to review this pre-therapeutic assessment and, if indicated, initiate treatment at W2.
In both clusters, patients refusing ART will be offered the opportunity to remain in care and will be asked to consent to six-monthly clinical assessment: describe (see section 6.4.4.3).
In both clusters, participants already established on ART from the Hlabisa HIV Treatment and Care Programme will be encouraged to be included in the trial and to transfer their care to the trial clinics. Participants already established on ART from private/other HIV treatment providers will be encouraged to take part in the trial monitoring procedures. They will be offered the option of 1) transferring follow-up to the trial clinic (the same follow-up procedures as for patients newly established on ART will then apply) or 2) continuing follow-up from their normal provider, without any discriminatory measures applied to them. During the home visit, all HIV-infected participants will be asked to give consent for their clinical records to be obtained from their normal provider should they choose not to attend the trial clinics. Information accessed would include ART status, ART regimen if on ART, CD4 counts and viral loads. This information will be used to estimate the ART treatment coverage in each cluster as this is directly linked to HIV transmissions. The nurse at each trial unit will liaise weekly with the Hlabisa HIV treatment programme clinics to ensure ease of information collection for patients who received clinical care outside of the trial clinics.
* Note on CD4 point of care: We will use the Alere PIMA™ device tool (http://pimatest.com/en/home.html) to measure CD4 in the clinics. The system has built-in quality control mechanisms, in addition to which the trial will develop its own quality control procedures (in association with CD4 laboratory, NHLS). Quality control will be performed on a 10% of participants selected by systematic random sampling using the enrolment log of each clinic as a sampling frame. All subsequent samples from the identified individuals requiring haematology assessment will also be assessed for CD4 count at the NHLS laboratory at Hlabisa hospital.

ART initiation criteria
In the intervention clusters, there are no immunological criteria for starting ART; all HIV-infected participants are eligible for ART. All ART drugs that will be used in the trial are those recommended in the National Department of Health Adult HIV management guidelines (see Table 1).
The standard first-line drug regimen for HIV-infected participants will in both arms be the fixed drug combination (FDC) Tenofovir (TDF -245 mg) + Emtricitabine (FTC -200 mg) + Efavirenz (EFV -600 mg) or Atripla® once daily. Participants who transfer their care while on the generic formulation of the FDC Tenofovir (TDF -245 mg) + Emtricitabine (FTC -200 mg) + Efavirenz (EFV -600 mg) which is available in the government HIV treatment programmes will be able to continue on this but for logistic reasons could be switched to Atripla® as they would then be less dependent on supplies from the government (See Appendix 15.1).

Supply
Atripla® will be provided by MSD during the study period, All other drugs will be provided by the KwaZulu-Natal Provincial Department of Health according to the standard of care.

Drug handling and storage
Trial drugs provided and shipped by MSD will be held centrally at Hlabisa Hospital pharmacy, under the supervision of the trial pharmacist. Supplies will be delivered weekly to the treatment sites within the trial area and kept in secure storage facilities.
At each delivery, the pharmacist will record the trial patient identification code, the name of the drug, the number of the batch, the expiry date and the accountability on the Drug Accountability form.
All trial drugs will be labelled with the patient name as per South African standards, trial number, and dosing requirements.
Patients will be provided with 4-weekly supply of drugs initially. Participants who have undetectable viral loads and no ongoing medical problems will be allowed 8-weekly supply of drugs to decrease frequency of clinic attendance to save them time and transport costs.
Participants will be requested to return all empty bottles and to bring any bottles in use to their follow-up visits. Unused drug must be returned to the trial site if a participant withdraws from the trial. Unused drugs will be disposed of through the existing disposal mechanisms in place at Hlabisa Hospital pharmacy.

Patient follow-up
The follow-up procedures for HIV-infected participants will be the same in both intervention and control clusters.

ART initiation visit
ART initiation will take place ideally within two weeks of enrolment at the trial clinic, unless purposely delayed for clinical reasons (e.g. TB treatment).
The results of baseline investigations of all patients eligible for ART initiation and consent for treatment will be reviewed with the nurse. Nurses will initiate ART in uncomplicated cases. As per the Hlabisa HIV Treatment and Care Programme, there will be specific guidelines regarding clinician referral prior to ART initiation (e.g. for patients with persistent TB symptoms but negative sputum smear, and patients with abnormal renal or liver function). Appendix 15.5 presents a flowchart of the follow-up provided to HIV-infected participants in the intervention and control clusters.
For patients already on ART and transferred from private/other HIV treatment providers, their ART initiation date will be the date of their first visit to a TasP clinic (for the trial purposes, it will not reflect their duration on ART).

Treatment duration
Once the decision to initiate ART has been made, patients will receive ART and be followed-up within the trial for a maximum duration of four years depending on time of enrolment in trial (see Table 2).   Any patient deciding to permanently discontinue ART will be offered the opportunity to transfer to the DoH fixed clinics while remaining under active follow-up in the trial: participants will be tracked during the household HIV testing rounds, their patient records may be accessed (see section 6.4.1).
During the first phase, all HIV-infected participants will be followed-up between 1 and 24 months. At the end of the first phase, HIV-infected patients will either be transferred to the Hlabisa ART programme if the trial stops, or followed-up until the end of the second phase.

ART follow-up visits
■ Participants receiving ART Participants initiating ART will be reviewed by the trial nurse monthly when they collect their medication until they have two consecutive undetectable viral loads measured three months apart following which follow up will become 8-weekly in those with no on-going medical issues.

Clinical evaluation (4-weekly or 8 weekly)
The nurse will document weight and proceed with a review of current morbidity and illnesses/hospitalizations since last visit (CHE form).

Interview/Questionnaire (six-monthly)
HIV-infected participants, whether on treatment or not, will be administered the social science clinic-based survey and interviewed at M6, M12, and 6-monthly until M48 (counting from entry into care), regarding specific topics according to visits (see Table  3). The clinic-based survey will be divided in two parts:  one part administered by the ART counsellor who will ask non-sensitive questions such as disclosure or health expenditures: the Social science clinicbased counsellor-administered questionnaire (SCC);  one part administered by an independent interviewer who will ask sensitive questions (which are subject to social pressure and thus to social desirability bias) such as sexual relations or adherence: the Social science clinic-based independent interviewer-administered questionnaire (SCI).
Here below in Table 3 is presented the distribution of topics to be asked either by the counsellor or interviewer.  I  I  I  I  I  I  I  I   Stigma and discrimination  I  I  I  I  I  I  I  I   Satisfaction with care  I  I  I  I  I  I  I  I C: counsellor-administered questionnaire I: independent interviewer-administered questionnaire; * only for ART-treated patients.

Adherence monitoring
Both an independent interviewer and the ART counsellor (who is not responsible for the treatment delivery) will monitor adherence at each visit for all participants receiving ART. This will include three categories of adherence measurement tools: Sub-optimal adherence will prompt an individual session with the HIV counsellor and nurse. Persistent sub-optimal adherence (over three consecutive visits) will prompt referral to the trial clinician. Specific problems may warrant referral to members of the Hlabisa HIV Treatment and Care Programme multidisciplinary team (pharmacist, social worker, psychologist, dietician, and home-based carers).

Laboratory assessments (M3, M6, M12, 6-monthyl up to M48)
Participants initiating a Tenofovir-based ART regimen such as Atripla will require monthly urea, electrolytes and creatinine for the first 3 months, at 6 months and then 6 monthly.
Participants initiating a Zidovudine-based ART regimen will require monthly FBC for the first three months, at 6 months and then 6 monthly.
Other screening (as required): Sputum will be collected for TB investigation (GeneXpert, microscopy and culture), as appropriate following screening on clinical symptom  Interview/Questionnaire: HIV-infected participants will be administered the clinic-based questionnaires (SCC and SCI) and interviewed regarding specific topics according to visits (see Table 3).


Based on their immunological status, treatment initiation will be proposed according to the procedures (pre-treatment assessment, treatment initiation and follow-up) described in sections 6.4.1, 6.4.2, 6.4.3 and 6.4.4 above.   ■ Participants on ART who prefer to remain within the Hlabisa HIV treatment and care programme Participants on ART who, after visiting the trial clinic, would prefer to remain within the Hlabisa HIV treatment and care programme will be asked permission for their records to be accessed by the trial nurse as per trial protocol (trial nurses will attend the nearest fixed clinic once a week). During home visits, all participants would be asked to give consent for their HIV records to be accessed by the trial team. Information required will be ART regimen, CD4 count and viral load. Linkage to existing records will be done using South African identification number and date of birth.

Adherence support
The importance of long-term adherence will be emphasised in the pre-treatment period during the treatment literacy classes, as per standard DoH procedures. In addition, at each trial site a peer support group will operate for the duration of the trial and all participants will be encouraged to actively participate.
Further, the trial clinics are positioned at sites no more than 45 minutes walking away, which for most people will be considerably closer than the DoH clinics. HIV-infected participants will also face a considerably shorter waiting time, in relatively pleasant surroundings.

Trial treatment modifications
Trial treatment modifications will be done according to national guidelines and developed in specific trial guidelines.

Occurrence of new pregnancies
Pregnancy testing (urinary hCG) will be offered at monthly follow-up to all women of childbearing age receiving trial drugs. In line with updated national ART guidelines, Atripla® will be used for women of childbearing age and during pregnancy irrespective of trimester of pregnancy (29).

Toxicity and treatment failure
In case of toxicity or intolerance, single drug substitutions will be allowed, as per South African guidelines (29,33). The toxicities monitored will be renal dysfunction (TDF), liver function (NVP) and teratogenicity among pregnant women (EFV).
In case of treatment failure, switch to second-line therapy will be recommended (see appendix 15.1). VL and CD4 cell count will be performed 3 and 6 months after initiation and then 6-monthly for all participants on ART. The decision to switch to second-line ART will be taken on the basis of two consecutive HIV RNA measurements > 1 000 copies/ml at least two months apart (as per current South African national guidelines) (29, 33). Information from genotyping will also be made available in realtime to the trial clinician to guide the decision whether to switch regimens. The standard second-line regimen will be ZDV + 3TC + LPV/r.
Participants positive for HBsAg will continue TDF in their regimen should they need to switch to second-line therapy.
Trial participants fulfilling criteria for immunological and/or clinical failure in the absence of criteria for virological failure will be reviewed by the trial clinician before any switch in drug regimen. Second-line regimen will be prescribed following a confirmed virological failure while the patient is on therapy, in accordance with South African guidelines.

Care of patients at the end of the trial
Trial participants receiving ART at the end of the trial will be transferred into the Hlabisa HIV Treatment and Care Programme and will remain on the same drugs. The drugs will then be provided by the KwaZulu-Natal Department of Health, whether firstline or second-line drug regimens (see Appendix 15.3).

Concomitant therapies
 Isoniazid preventive therapy (isoniazid 300 mg orally od) will be provided to HIVinfected participants in the intervention and control clusters as per national and provincial policies at the time of the trial (implementation in progress).

Multivitamins (1 tablet od) supplementation will be provided to HIV-infected participants on ART in the intervention and control clusters as per South
African guidelines.

Sample handling and storage
Blood samples will be collected by trial field staff at the time of enrolment and by the nurse during subsequent visits to the trial clinics for HIV-infected trial participants according to the schedule presented table 1. The trial will utilise both the National Health Laboratory Service (NHLS) laboratory at Hlabisa Hospital and the Africa Centre laboratory at the University of KwaZulu-Natal, Nelson R. Mandela School of Medicine in Durban. Storage of samples will take place only at the Africa Centre laboratory. No samples will be transported out of South Africa without seeking appropriate permissions and approvals from the Ethics Committee, the University and South African regulators. Detailed procedures for sample collection, transport, processing, and storage will be included in a separate laboratory procedures manual/protocol.
The DBS samples collected during the HIV testing rounds will be transported daily via the sample processing unit at the Africa Centre to the Africa Centre laboratory in Durban. Two (2) of the DBS will be used on arrival at the laboratory for HIV testing.
The remaining DBS will be stored at -80°C in the Africa Centre Virology Laboratory in Durban.
Blood will also be collected by venepuncture:  Table 2.
A total of 30 ml will be drawn each time:  10 ml of blood will be sent to the NHLS laboratory for routine haematology, biochemistry, and HBV testing (these samples will not be stored)  20 ml of blood (2×10 ml tubes) will be sent to the Africa Centre laboratory for plasma viral load testing and storage at -80°C. Table 5 summarises the definition of each trial outcome. If not explicitly mentioned, these outcomes will be documented in both phases of the trial, within the 10 clusters for the first phase and within 22 clusters during the overall trial.  Table 6 presents the outcomes that will be measured within the planned sub-studies if funded.

Primary outcome
HIV incidence will be measured after multiple HIV testing periods of 24 months in 12 clusters, 36 months in 6 clusters and 48 months in 4 clusters depending on the person years of observation in individuals initially identified to be HIV negative.

HIV status measurement, using DBS with longitudinal follow-up
HIV testing will be longitudinally linked at the level of the participant and be conducted in six-monthly intervals. HIV testing will follow the standard South Africa DBS protocols used in the Africa Centre population-based HIV surveillance in past years (34). For the first phase of the trial, the duration of the three testing rounds will be six, then four and four months for the first 4 clusters and 6-monthly for the additional 6 clusters. For the second phase of the trial, four rounds of 6-monthly HIV testing will be conducted in all 10 phase 1 clusters and the additional 12 clusters in phase 2 over a period of 24 months.
Serologic status will be determined by antibody testing with a broad-based HIV-1/HIV-2 ELISA (Vironostika® HIV-1 Microelisa System (Biomérieux, Durham, NC, USA) followed by a confirmatory ELISA if the first test is positive (Wellcozyme HIV 1+2 GACELISA; Murex Diagnostics Benelux B.V., Breukelen, The Netherlands). In cases of discordant results a third ELISA test will be carried out.

A locally validated test for recent infection: capture BED enzyme-linked immunoassay (cBED assay)
For HIV-positive DBS a cBED assay will be performed during the first phase only to establish whether the HIV infection is recent (less than 6 months) or non-recent and confirm incidence estimates.
The addition of a test of recent infection to the standard longitudinal HIV testing serves three purposes, it will:  Allow an immediate assessment of HIV incidence in the trial population at baseline (to confirm the assumptions made re sample size; without the need for a second longitudinal assessment).
 Improve the precision of the overall longitudinal measurement.
 Allow inclusion of individuals who refuse repeated testing but consent to testing once in the sample for incidence estimation and thus increase the power of the trial.
The cBED assay has been locally calibrated (35) and is thus an appropriate measure of HIV incidence at baseline for a community-randomized trial. Moreover, some of the calibration problems encountered in applications of the BED assay in other settings (36), become irrelevant in repeated application of the BED assay to estimate HIV incidence differences over time, since it is reasonable to assume that calibration parameters remain constant over time intervals of a few years.
While it would thus theoretically be possible to rely solely on BED assay testing in cross-sections of the population, which are not longitudinally linked at the individual level, such a measurement strategy does not seem advisable. The BED assay has not yet achieved acceptance as a routine approach to measure HIV incidence. Relying solely on the BED assay could thus reduce the perceived strength of evidence emanating from the TasP trial. However, we will use the cBED assay to identify people with likely recent seroconversion in the first round.

Acceptability and uptake of HIV testing
Acceptability and uptake of HIV testing, current knowledge of HIV status, recent HIV testing history will be calculated for each successive rounds of home-based HIV testing and will be based on data collected during the home visit.
Acceptability of HIV testing will be estimated among different population subgroups, and will be defined among others as:  the proportion of participants who are tested for HIV among all those eligible (individuals identified though the household registration)  the proportion of participants who are tested for HIV among those not previously tested  the proportion of participants who are tested HIV-negative at first round and who return to be tested once or more during the subsequent testing rounds. In addition, the uptake of HIV testing estimations will be refined using the data from trial clinics and the DoH clinics on persons who attend for VCT. Participants from the trial sites who attend either service with their referral cards, given out during the rounds of home-based testing. Their data can be added into the total to provide estimates of testing coverage and the acceptability of different testing modalities. These outcomes will be calculated for successive rounds of testing. Essential socio-demographic data on nonresponders (i.e. those declining HIV testing and the questionnaires and not subsequently attending either trial or DoH clinics for VCT) will be collected in each round and can in turn be linked in order to estimate an overall uptake rate over individual and over successive rounds of home-based HIV VCT. It will also be possible to estimate the appropriateness of the time intervals between testing rounds to maximise uptake.
Other proxy acceptability indicators can be calculated including uptake of HIV treatment and care among those testing HIV-positive either in the home-based HIV VCT or the local clinics (trial or DoH). We will also have individual level data on participants' HIV-testing histories, disclosure of HIV status, perceptions and expectations of treatment and collected at initial enrolment into treatment and care. It will be possible to calculate these estimates for each individual round of testing and cumulatively over the trial period.

Behavioural changes at individual level and sexual partnerships
Measurement of sexual partnerships, relationship status, prevention behaviours (e.g. condom use and HIV testing outside the trial), circumcision status, changes in contraceptive usage, attitudes towards PLWHA, perception and knowledge of HIV treatment, will be collected from all enumerated participants during the successive rounds of home-based testing. This will facilitate longitudinal as well as repeat crosssectional analyses regardless of participants' self-reported HIV status.

Quality of life
Quality of life will be explored:  At cluster-level, among all participants responding to the IQ questionnaires during the first and last rounds of home-based HIV testing during the first phase. The EQ-5D scale will be used for this purpose since this instrument is a short generic health instrument which can be administered to both HIVinfected people and HIV-negative people or people of unknown HIV status.
 Among all HIV-infected participants attending the study clinics. These measurements are planned for at W0 (i.e. date of entry into care for treatment ineligible patients and any of the pre-treatment visits for ART-eligible participants), M6, M12, and 6-monthly up to M48 (see Table 2, Table 3 and Table  4). The quality of life module will be designed using standardised validated instruments appropriate to the context and culture of the participant population: properties have been evaluated and validated in different context including Sub-Saharan Africa and the instrument has been shown to be sensitive to differences in culture, gender and ethnicity (38). As the PROQOL-HIV has not been validated as an isiZulu language instrument, a short linguistic validation study could be implemented (to be described at a later stage)  the HIV/AIDS stigma instrument for PLWHA (HASI-P): this tool assesses stigma perceived in PLWHA, which is one of the dimensions of the quality of life. This instrument has been built to measure perceived stigma by PLWHA in Southern African countries and its psychometric properties have been validated in IsiZulu language (39). This face-toface questionnaire includes 33 items regarding the occurrence of different events, which may have happen because of the HIV status (e.g. ask to not touch child, blamed for HIV status, called bad name) as well as thoughts or feelings.

Cost-effectiveness performance
A cost-effectiveness analysis (CEA) will be carried out based on data from the TasP trial and will be completed with a mathematical (Markov) model to estimate the effectiveness and costs of the strategies compared over time, i.e. comparing treatment initiation as soon as HIV infection is diagnosed versus delayed treatment initiation according to WHO recommendations.
The main outcomes of this CEA will be the number of life years gained and also the number of Quality Adjusted Life Years (QALYs) saved.
A generic health instrument is required to determine different healthcare states which can be classified according to preferences. We will use the EQ-5D scale, which has been widely used in the literature to conduct cost-utility evaluation, both in Northern and in Southern countries, including South Africa (40,41). This instrument has been found to discriminate between different health states amongst PLWHA and has been validated in isiZulu formats (42,43).
Both direct costs, as well as indirect costs will be evaluated both at the cluster level and among HIV-infected participants. The questions for the CEA will be integrated into the IQ at the cluster level within the CRF and the clinic-based questionnaires to be administered to HIV-infected participants. The following costs will be collected throughout the trial period at different time periods:  Table 2, Table 3 and Table 4).

Community awareness, attitudes and behaviours
Community awareness, attitudes and behaviours will be extensively studied in the first phase. These are measured at the individual level in the IQ questionnaires completed during each successive round of home-based testing. Other global indicators will be derived from participation rates in successive rounds of home-based testing. The extent to which the same methodology or a more complex one will be used in the second phase will be decided at the end of the first phase.

Societal response
Societal response is a composite outcome, which will be assessed during the second phase of the trial only, using a variety of methods and indicators: With these various outcomes we will be able to compile an extensive array of measures that can be combined and compared between clusters and between individuals (HIVinfected and non-infected). This will in turn be supplemented with qualitative data, which will provide a nuance image of the overall social impact of the programme and the trial.

Acceptability and uptake of entry in care and treatment
Using the patient identifiers and mechanisms already described we will be able to monitor acceptability and uptake of treatment at several levels.
At the individual level we will be able to estimate the time from receipt of HIV test result to enrolment into treatment at the trial clinics or DoH fixed clinics (if any participant chooses that option). We will be able to monitor the transfer of care of patients currently being monitored (as yet ineligible for ART) in the fixed clinics to the trial clinics. The employment of a tracker will ensure that people eligible for treatment in the trial clinics who do not attend, and those that subsequently default from care, are followed-up, are appropriately assisted to access the trial clinics, if necessary, and that they are not otherwise lost to follow-up.
In addition to measurement of these key public health outcomes, with the data collected at treatment enrolment it will be possible to characterise in detail the population entering into treatment and care and to estimate social factors that facilitate entry (e.g. household composition, social support, disclosure to family, treatment knowledge). All of these intermediate measures will be useful in determining the need and content of any special interventions to support treatment uptake in the main trial.

Therapeutic success/evaluation of ART
This will be estimated as the proportion of HIV-infected participants who have undetectable VL after six months of ART.

Programme retention
Retention in care will be assessed at 12 months, 24 months and 36 months postenrolment. Retention will be defined as those still under active follow-up in the trial. Loss to follow-up on ART will be defined as ≥ 3 consecutive missed appointments. For those not eligible for ART, loss to follow-up in control clusters will be defined as > 9 months from last clinic visit and/or CD4 cell count.

Mortality, severe morbidity, and tuberculosis
Secondary endpoints to be compared between the intervention and control clusters include all-cause mortality, HIV-related mortality, WHO stage IV disease, serious non-AIDS events, and tuberculosis. Serious non-AIDS events include cardiovascular disease (stroke, myocardial infarction), end-stage renal disease, decompensated liver disease, and non-AIDS-defining cancers. Tuberculosis includes both pulmonary and extrapulmonary disease, either with or without microbiological confirmation.

Adherence
Adherence will be measured every 4 to 8 weeks depending on frequency of clinic attendance for patients on ART using a scale that has already been translated and validated by Africa Centre researchers with patients from the Hlabisa Treatment and Care Programme where previous analysis has focused on consistency of reported nonadherence (44). This tool, using visual analogue scale, pill identification test and pill count, will be included in the CFU questionnaires.
However, as non-adherence was rarely reported and as the questions used performed poorly in identifying patients with treatment failure (44), we will test an additional scale constructed to limit both recall and social desirability bias and which has been tested in different settings (45)(46)(47). This tool includes several questions related to dose taking during the previous 4 days and the respect of the dosing time schedule during the previous 4 weeks. Adherence scores which are computed using a validated algorithm allowing to classify patients into highly adherent, moderately adherent and lowly adherent have been found to be significantly associated with viral load (46). Another item focusing on the occurrence of treatment interruptions lasting more than 2 consecutive days during the previous 4 weeks was found to be a predictor of resistance development (48) and has already been tested in another context (45). This second measure of ART adherence will be surveyed every 6 months (i.e. at M6, M12, and 6monthly up to M48) and included in the Social science clinic-based intervieweradministered questionnaire (SCI).

Paediatric HIV infection
The sub-district has a well-functioning PMTCT programme, which includes HIV DNA testing (PCR on a DBS) of infants at six weeks of age, repeated one month after cessation of all breastfeeding. A PMTCT database has been developed in the subdistrict, charting the progress of pregnant women from the time of antenatal booking through to infant HIV testing after cessation of all breastfeeding. It has been shown in this programme that ART introduction in mothers and infants may already have had more impact on early life mortality than PMTCT itself (49).
In the trial intervention clusters all HIV-infected women will be initiated on ART irrespective of the CD4 count or the clinical stage. In the control clusters, women who are not yet eligible for HAART (i.e CD4>350 cells/µL) will be managed according to the best available PMTCT guidelines approved by the South African national authorities (Currently WHO option B) (29). This involves the prescription of FDC triple therapy (TDF+FTC+EFV) to pregnant women and breastfeeding mothers with CD4>350 until one week after the complete cessation of breastfeeding. This is adapted in certain situations such as the presence of active psychiatric illness or renal disease. Women receiving TDF/FTC/EFV for prophylaxis rather than for their own health will undergo, WHO clinical staging and hepatitis B testing prior to discontinuation of ART. If a woman treated with TDF/FTC/EFV has stage 3 or 4 disease or evidence of hepatitis B infection (hepatitis B surface antigen positive) then the regimen will not be stopped after the cessation of breastfeeding but will continue lifelong. CD4 count will be done 6 months after the cessation of breastfeeding or if the last CD4 count was done more than 12 months ago.
Women in both the intervention and control clusters will be asked about pregnancy at each of their routine visits; if pregnant, data will be collected as part of the trial including date of delivery, mode of delivery, stillbirth, premature termination, infant outcome, and infant DNA status at 6 weeks and after cessation of all breastfeeding if known, as antenatal care, delivery and post-natal care are managed in the DoH clinics.

HIV drug resistance
HIV-1 genotypic resistance testing will be performed for: Programme. Results will be returned to the trial clinician within 14 days of sample collection. Genotyping for all participants who seroconvert within the trial -these results will not be available to clinicians in real-time.
The pre-treatment samples will be used to assess the surveillance of transmitted drug resistance during the trial whereas virological failure genotypes will be used to estimate the level of developed resistance mutations. Sequences will be produced using the ANRS guidelines for genotyping and analysed using drug resistance algorithms, such as the Stanford HIVdb, ANRS and Rega algorithms.
Further, given the short time of the trial and potency of the ARV drugs to be used in the trial, in a sub-study (n~10) we will also consider genotyping at the DNA level of participants with undetectable VL while on ART to assess development of resistance at low viral replication levels and sequencing minority populations using either direct PCR or pyrosequencing methods in order to study the link between minority populations and clinical outcome.

Molecular epidemiology
We will analyse all sequences using a phylodynamics framework. Phylodynamics can be used in order to estimate the date of origin of epidemiologically important events such as the introduction of a new viral strain in a geographical area and identification of transmission networks between intervention and control clusters.

Data collection tools
A number of detailed standard operating procedures (SOPs) will be developed for the counsellors involved in the home-based testing and surveillance, the trial clinic nurses and counsellors, the physicians, the pharmacy and the laboratory technicians. A combination of quantitative and qualitative trial instruments will be used to assess key intermediate process, behavioural and outcome indicators in the trial.

Household and individual home-based questionnaires
Throughout the trial (first and second phase), in each round of home-based HIV testing, we will collect basic information using home-based questionnaires at household and individual level.
In the first round, the TasP household registration questionnaire (THR) will be administered to the head of household to document household-level social and demographic characteristics: number and age of residents. Housing characteristics (water, electricity), household income, food security among others will be documented using the TasP household information assets questionnaire (HHI). These household questionnaires will be repeated to allow for the updating of information about household members at each round in order to document changes in household.
During the first phase of the trial, the TasP home-based individual questionnaires (IQ1 to IQ3) will be composed of a core questions repeated at each round and a set additional questions specific to certain rounds: individual-level social and demographic characteristics, HIV testing behaviour, sexual behaviour, partnership and sexual network patterns, attitudes and beliefs about HIV infection, HIV testing and treatment, stigma and disclosure, healthcare use and healthcare expenditures and changes within the households, and quality of life (see Table 7).
In phase 2 of the trial, identical home-based individual questionnaires designated IQ will be administered during the six-monthly rounds of HIV testing. The phase 2 IQ will address all components listed in Table 7.

Social science Clinic-based survey
The clinic-based survey exploring HIV-infected participants' behaviours and socioeconomic status will be composed of different sets of questions administered repeatedly every six months or specifically at baseline (i.e. date of entry into care for non-eligible patients and date of any of the pre-treatment visits for ART-eligible patients), as presented in Table 3. The topics explored will be: testing experience (baseline only), ART knowledge, ART perception and decision, disclosure, sexual behaviour, selfreported ART adherence, HIV Quality of Life, stigma and discrimination, social and community support, economic and health expenditures, satisfaction with care.
The clinic-based survey will be divided in three questionnaires:  the Social science Clinic-based Baseline questionnaire (SCB), administered by the ART counsellor at baseline;  the Social science Clinic-based Counsellor-administered questionnaire (SCC) on non-sensitive questions such as disclosure or health expenditures, administered each six months of follow-up and  the Social science Clinic-based independent Interviewer-administered questionnaire (SCI) with sensitive questions (which are subject to social pressure and thus to social desirability bias) such as sexual relations or adherence, administered each six months of follow-up.

Case Report Forms (CRF)
The case report forms will be completed by the counsellors and the research nurses in the trial clinics. There will be baseline (CBC) and follow up (CFU) CRFs to be completed by the counsellors and a combined history and clinical examination (CHE) CRF to be completed by the research nurses. The same CRFs will be completed for all participants irrespective of their ART status at enrolment into the trial (See Table 5).

Clinic Activity Reports (CAR)
Throughout the trial, each trial clinic will keep detailed monthly activity records. In addition in phase 2 we will attempt to audit patient waiting times, staffing levels, stock outage, the adequacy of trial logistics and support using tools developed for the task.

Qualitative data
In the first phase only, two qualitative studies are planned to address acceptability of repeat testing, adherence and quality of life among those starting treatment in the trial clinics.
In the first of these we will convene a consumer advisory panel (CAP) in each of the four clusters. This panel will work with the social science team and the trial's community liaison officer to provide an on-going monitoring of the community experiences of the trial, perceptions and understanding of the intervention and early advice on any possible problems. The members of each panel will also serve as key informants and expert advisors to help the trial team ensure that community entry, awareness and education plans are fully developed for the second phase of the trial. The four community consumer panels will be conveyed according to a purposive sampling framework. Africa Centre researchers are currently using this methodology to develop and deliver youth HIV prevention interventions. In this instance we will assemble and engage repeatedly with the four purposively selected groups on a bi-monthly basis to discuss specific issues, which we believe the panel members, shared in common. Our community consumer panels will comprise four purposively selected groups from among the different communities (two members from each community). The four panels will be comprised of women, men, elderly, health care workers, and traditional authorities. Each panel will meet on six separate occasions with the same facilitator. All sessions will be video-recorded with participants' permission. Panel meetings will adopt a combination of different approaches including conventional focus group discussions and participatory action methods including joint problem solving and assumption of expert advisor roles.
The second small qualitative study will use repeat in-depth interviews (IDI, n=15) to explore the adherence and quality of life in HIV-positive people initiating treatment in the trial clinics. This is a unique opportunity to examine adherence and quality of life qualitatively, and in the context of people who would not normally be receiving treatment. It also offers an important opportunity to explore understandings of the importance of adherence in the context of both personal health benefits and HIV prevention.

Survey of health care professionals
The implementation of universal testing and treat immediately intervention will represent major challenges for the health care system and especially for human resources which are one of the key factors for the success of the intervention. Sub-Saharan Africa countries currently encounter lack of qualified health care professionals, unequal geographical distribution of health resources, generalized low wages, difficult working conditions and lack of carrier development (50,51). Such factors have been shown to have an impact on human performance and may jeopardize quality of care, as well as HIV treatment delivery (52).
To identify potential obstacles to the implementation of TasP in HIV care, treatment and prevention services and to identify changes needed to improve the adherence to TasP principles by service delivery personnel, a quantitative survey will be conducted among health care providers in charge of PLWHAs and working in the facilities included in the TasP trial, both the trial clinics and the DoH fixed clinics. Data will be collected using a quantitative survey instrument previously used in a research programme about HIV care among medical professionals in Cameroon (53). Data collected will include information on socio-demographic characteristics, training and experience in HIV care, working conditions, practices and knowledge about HIV and ART management, opinions about the TasP intervention. This survey will be carried out once separate from the routine care and in the second phase of the trial.
In addition, data relating to the characteristics of the health care facilities where TasP will take place, such as types of health services, size of HIV clientele, number of ARTtreated patients, human resources in charge of HIV care, working time devoted to the care of PLWHAs and staff compensation to estimate the cost of human resources involved in patient care, shall be obtained through access to institutional reports, computer systems and interviews of each health centre's managers and staff. This second survey will be carried out at baseline and at 12-month, both in the second phase of the trial.

Budget impact analysis
A budget impact analysis (BIA) will be planned in the first and second phase of the trial to examine the potential financial impact, at macro level, that is on national, regional or local health budgets, of the introduction of the TasP strategy into the healthcare system. This approach will be a complement of the CEA but does not include additional data collection. The impact on the total costs will be estimated. A variety of sensitivity analyses will be carried out varying for example the level of coverage, the rate of switching to second-line regimens, or drug prices.

Methodological and statistical considerations 8.1 Cluster selection
The randomisation units within the TasP trial will be clusters within Hlabisa subdistrict outside of the DSA (see Figure 1). Randomisation will occur at the start of the first phase.
The area consists of 150 local areas (neighbourhoods). These were aggregated into 34 clusters of between one and six contiguous neighbourhoods, each cluster comprising an average of 1 000 individuals >15 years of age. Clusters were designed to encompass social networks based on earlier studies in the DSA (54), to keep the potential for crossarm contamination to a minimum. To further reduce the potential for contamination, we ensured that all resulting clusters were of significant geographical size (>15km 2 in area). This meant that the number of individuals in peri-urban clusters were larger than their rural counterparts.
Adjacent communities representing relatively distinct social and sexual networks 1 were grouped together to form the units of randomization (median population ≥ 15 = 1 000; median area = 19 km 2 ). From a statistical efficiency perspective (least sample size) more clusters of fewer people should be used. However, having very small units of randomization would result in a large potential for contamination. With this trade-off between statistical efficiency and contamination potential in mind we combined adjacent neighbourhoods together to form relatively large randomization units (with similar numbers of participants) with a distinct social identity. This considerably reduces the risk of contamination whilst still retaining a relatively large number of clusters for the trial (56).
To further evaluate the potential impact of contamination of the results on the outcome of the trial we have run extensive geographical simulations of possible randomization scenarios. The results of these analyses demonstrate that chance groupings of neighbouring communities randomized to the same arm of the trial result in 'superclusters' with a median of aprooximately 4 000 participants ≥ 15 years of age and median size of 48 km 2 . Empirical data from a sexual behaviour survey conducted in 2005 showed that 55% of men interviewed reported that their most recent partner in the past year lived in the same immediate izigodi (a Zulu term for an administrative area headed by a local chief), which have a median area of 16.9 km 2 . Extrapolating from these results, we are expect that only a very small proportion of trial participants will have sexual partners outside these 'super-clusters' (median size=48 km 2 ). The exceptions to this rule will be people residing near the borders of these 'super-clusters' and the chance occurrence of singleton communities in the randomization.
We considered trying and further reducing the small amount of inter-arm contamination by the designation of a series of 'buffer' communities around each community used in the trial. However, this was rejected for two reasons: a) It would have the undesirable effect of reducing the size of the 'super-clusters' and hence increase the chance of having partners outside the area to which the participant was randomized and b) It would increase the risk of an HIV-negative participant in an intervention community of having an HIV-positive untreated partner in the neighbouring buffer communities (because the level of ART coverage in buffer communities would be even lower than control communities where ART coverage will increase through increased access to VCT services delivered as part of the trial).

Principles of the model
We used STDSIM to provide estimates of the potential impact of the TasP trial. STDSIM is a well-established stochastic microsimulation model of the spread and control of HIV (57)(58)(59)(60)(61)(62)(63)(64)(65)(66)(67)(68)(69)(70)(71). The model that simulates the life course of individuals in a dynamic network of sexual contacts and consists of four modules: demography, sexual behaviour, transmission and natural history, and interventions. The demography module covers the processes of birth, death, and migration. Processes for initiation and dissolution of sexual relationships, mixing according to age preference, sexual contacts within relationships, and sexual contacts between female sex workers and their male clients are defined in the sexual behaviour module. In the transmission and natural history module, transmission probabilities per sexual contact are specified for HIV and 5 other STDs. The interventions module specifies the timing and effectiveness of control measures in curbing transmission (e.g. condom use) or enhancing survival (e.g. antiretroviral therapy). STDSIM has been extensively used to evaluate behavioural interventions (60)(61)(62)72), syndromic treatment for STIs (63,69), male circumcision (70), explain different HIV epidemics in sub-Saharan Africa (64), and, more recently, the impact of ART on HIV epidemics (57)(58)(59)73). For the impact estimates of the TasP trial we used the previously published quantification for rural KwaZulu-Natal, South Africa. The model was built using data from the Africa Centre and accurately captures demographic, behavioural, and epidemiological characteristics of the local epidemic (58, 59, 73). Table 8 gives an overview of the main parameters and assumptions related to the trial impact estimates. We simulated an intervention of bi-annual screening of the population aged 15 and over, and compared the impact of immediate ART for all HIV infected individuals (treatment-arm) with ART only for those with CD4 cell counts of ≤350 cells/µL. We ran the model separately for all clusters in order to capture the phased roll-out design of the trial. We assumed the intervention to be rolled-out in a total of 22 clusters: four clusters start in 2012, six in 2013, and a further 12 in 2014. We examined the impact on incidence rates, cumulative incidence, and ART coverage up to end-2016. The parameter settings used are those that gave a 30% reduction in cumulative incidence.  (58,59,73).

Model input parameters
Our model predictions give an estimated reduction of cumulative incidence in end-2016 of 31.7%, or 28.8% by end-2015 (Table 9). Incidence rates in the control arm would decline from 2.4/100 PY in 2012 to 1.4/100 PY at the end of 2016. In the intervention arm, incidence would decline from 2.4/100 PY to 0.7/100 PY over the same period (see Figure 5). Mid-2015 we will be in a strong position of conduct an interim analysis to show whether a significant reduction in incidence could be detected either at the end of 2015 or 2016. The details of this analysis are described in detail below. Table 9 Projected 5-year impact of cluster-randomized trial of treatment as prevention on HIV incidence and ART coverage in rural KwaZulu-Natal, South Africa.

Average incidence over 5 year period (per 100 PY)
Control-arm n/a n/a n/a 1.7 Treatment-arm n/a n/a n/a 1.2

Sample size calculation based on model assumptions
The initial design proposed a cluster-randomised trial with 2x17 clusters, covering a total population of 34,000 adults aged 16 years and above followed over two years. However, a phased approach of trial implementation was adopted as feasibility and acceptability were deemed critical for the overall evaluation of the intervention. In this design at the end of 2015, four clusters would be followed up for four years, six clusters for three years and a further 12 clusters for two years. This would give a total of 22 clusters, 58 cluster-years of follow up for a total trial size of 22,000.

■ Primary Outcome
Our sample size calculations are based on an overall 34% reduction in cumulative HIV incidence in HIV-negative participants over 58 cluster-years of observation (see Table  10). To have sufficient numbers to be able to demonstrate this difference, our sample size calculations indicate that 22 clusters (11 in each arm), with 1,000 participants >15 years of age in each cluster (N=22,000; 17,600 HIV-negative), are required to achieve this objective. Underlying statistical parameters are as follows: a 80% power to detect this difference, an alpha-type-one error of 5% (two-tailed) and an allowance of 20% of study participants lost to follow up. We assume also a coefficient of variation of 0.25 to account for within-group variation between clusters. In the control arm we expect an average HIV incidence of 2.25 per 100 person-years over the course of the trial (this is 17% lower than the incidence currently observed in the Africa Centre DSA and recognises that incidence will likely decline in the overall population as a result of public sector roll-out of ART according to South African guidelines).
In the intervention arm we expect an average HIV incidence that is 34% lower than the control arm i.e. 1.485 per 100 person-years.
To take into account the different follow up times in the different groups of clusters opened at different times, we based our calculations simply on the proportion seroconverting in the intervention versus control clusters over the course of the trial. This approach will yield appropriately conservative sample size estimates.
In each cluster we expect to register 800 HIV-negative individuals of whom we expect to be able to successfully follow up 640. The average follow-up time for each individual is derived by subtracting half a year from the follow-up time in the cluster in which they reside. So for example, in the two control clusters that are followed up from 2012 to 2015, the total person-years of follow up = 640*2*3.5 years =4,480 which at an annual incidence of 2.25 per 100 person-years would equate to 100.8 sero-conversions observed by 2015.
Overall, the total person-years of follow up in each arm of the trial will be an estimated 15,040 (and 30, 080 person-years of observation in the whole trial). Thus, at an annual incidence of 2.25%, 4.8% of those observed to be HIV-negative at baseline would have been observed to seroconvert in the 11 control clusters by the end of 2015 that is 338 new HIV infected. The corresponding figure for the 11 intervention clusters would be 3.1725% for around 223 new HIV infected. Therefore, the number of HIV infected patients in the 11 intervention clusters will be 2098 (338+1760) and 1983 in the 11 control clusters (see Table 11). (Note that this method, using average follow up time provides very similar total person-years as in the approach accounting for different follow up times in different clusters.) Within the overall trial, two kinds of secondary objectives are distinguished: those measured on the total population (acceptability and uptake of HIV testing, behavioural changes…) and those measured in HIV-infected patients only (morbidity and mortality, virological response and resistance). We describe below some of our expectations regarding the latter.

Mortality rates in HIV-infected patients
Assuming a 9 per 100 person-years rate of early mortality in the control clusters based on published data from the Hlabisa HIV treatment and care programme (32) for the post-2011 CD4 350 eligibility threshold, and assuming that the median duration of follow-up while being infected is two years for prevalent cases and one year for incident cases, HIV infected patients will contribute 3858 person-years in the 11 control clusters and 3743 person-years in the 11 intervention clusters. Following these assumptions, we will be able to show a reduction of at least 33% (incidence of 6 per 100 py) in the intervention clusters.
If we were to use only data from phase 1 on the 5+5 clusters, then we would be able to detect a reduction of 50% in the rate of early mortality (from 9 to 4.5 per 100 personyears of follow up) only.

Proportions in HIV-infected patients (adherence, virological success, resistance)
Assuming about 182 HIV-infected patients per cluster (around 4000 infected patients in total) we will be able to show, with a statistical power of 80%, at least a 17% difference for characteristics with a prevalence of 60% or 22% for a prevalence of 80% or 25% for a prevalence of 90% (that is for example, an adherence of 90% in control vs. 65% in intervention clusters).
If we were to use only data from phase 1 on 5+5 clusters, then we would be able to detect a difference in the proportions by cluster arm in a variable that has a 90% prevalence of 37% (compared to the 25% noted above for the 22 clusters).
Calculations were performed according to Hayes & Bennett formulas for rates and proportions assuming 20% of loss to follow-up, 5% type I error.

Randomization
Randomisation was performed by the trial statisticians (F. Tanser, R. Thiebaut) at the start of the trial for all 34 clusters. Random numbers were generated for all 34 communities in the trial area using MapInfo 10.0. Communities were randomly allocated in equal measure to control and intervention communities (17:17). To minimise the degree of between-cluster variation, communities were stratified on the basis of predicted HIV prevalence. Randomisation was carried out within each stratum to derive an equal number of control and intervention communities per stratum.

Criteria for continuation/discontinuation at the end of the first phase
Given our proposed phased approach, interim analyses will be conducted at three possible time-points over the course of the trial. As well as analysing levels of uptake of HIV testing and coverage of treatment, we will analyse differences in cumulative HIV incidence between intervention and control communities at the end of 2013 (≈≈19% of total person-time of observation accrued) and 2014 (≈≈53% of total person-time of observation accrued). Our modelling results demonstrate that any initial reduction in HIV incidence is likely to be relatively modest given the expected non-linear decline in incidence overtime (74). However, the trial will be discontinued if the interim analyses provide clear evidence of sexual disinhibition or reduced treatment adherence in the intervention clusters relative to the control clusters, that is, that the hazard ratio of acquiring new infection (intervention versus control) is >1.0 (p<0.05). In 2014 we will conduct a predictive power analysis to quantify the likelihood of observing a significant reduction in incidence if any observed reduction is sustained to the end of 2015. The predictive power probabilities will be calculated using the available information and several prior distributions. Futility will be concluded if the predictive power probabilities are < 0.2 or efficacy if > 0.8.

Feasibility
Clear indication that given the parameters measured during the first phase (HIV prevalence in the clusters and assessment of incidence at baseline on the basis of cBED, initial HIV testing uptake, repeated HIV testing uptake, ART treatment uptake, migration and the extent of sexual partnerships with people outside the trial setting) the TasP trial will lack the statistical power to detect significant HIV incidence differences between the intervention and control communities. Such observations will lead to the conclusion that the trial is not feasible.
The criteria used to decide on continuation of the trial after the first phase are those that substantially affected the impact of the intervention in mathematical modelling simulations: prevalence of undiagnosed HIV infection before the trial, CD4 level of HIVinfected participants who are undiagnosed or diagnosed but not yet on ART before the trial, ART initiation criteria in the control arm, internal migration, test acceptance and the linkage to care. The thresholds are defined according to the results of the mathematical model (see section 8.2).

Acceptability
Clear indication from the clinic-based survey (section 7.4.2) and the qualitative in-depth interviews conducted during the first phase that the TasP approach is not acceptable in our setting. Acceptability will be assessed using a combination of quantitative and qualitative analyses of community attitudes and beliefs about HIV, HIV VCT, stigma and disclosure, participation in TasP, and the acceptability of ART for the benefit of other community members.
The DSMB will, in addition to the criteria indicated in the protocol, be able to consider additional parameters measured during the first phase such as tuberculosis incidence.
A complete "pause" between the TasP first and the full trial phase is not feasible logistically, because it would imply discontinuation of structures built during the first phase in the intervention and control communities which would need to be utilized during the full trial phase. However, detailed data analyses and interpretation will take place in half-yearly intervals throughout the first phase, ensuring that TasP trial procedures are adopted in real-time to fully take into account the lessons learned during the first phase.

Statistical analysis
An intention-to-treat analysis, assuming sexual partnerships occur within clusters, based on the randomisation clusters, will be used for the primary trial outcome (incidence of HIV-1 infection). Incidence rates per 100 person-years will be calculated for the whole follow-up period and compared between the intervention and control arm. Cluster randomized trials require more a complex analysis than that for individual randomised controlled trials (75). Observations on participants in the same cluster tend to be correlated; therefore it is imperative that the intra-cluster variation must be accounted for during the analysis of the trial. If this correlation is ignored in the analysis and the same techniques are employed as for individual randomised controlled trials, the associated variance of the estimate would be underestimated and lead to unduly narrow confidence intervals. To ensure valid estimates, adjusted incidence rate ratios in the intervention group relative to the control group will be based on a multilevel Poisson regression taking into account the intra-cluster correlation and the repeated measurements when needed.
In detail: we will perform analyses using cluster-level approaches as robustness analyses for checking consistency of the results. However, we will focus on Poisson regression (including random effects) for the main outcome (HIV incidence) and logistic regression with generalized estimated equation (GEE) approach for binary outcomes, thus allowing for both levels (individual and cluster levels) with easy adjustment for confounding factors (76). As the method does not perform well for a low number of clusters, we plan to use the correction (of the sandwich estimator and statistics distribution i.e. t-distribution) proposed by Mancl & Derouen (77) which is particularly appropriate when the cluster size does not vary.
Because of the possibility that the randomisation may lead to some imbalance in the distribution of risk factors across the trial arms, an adjustment for baseline characteristics known to be associated with HIV transmission risk in this population will be undertaken. These include HIV prevalence of the cluster, age, sex, marital status, education level attained and migration status.
In addition, the extensive data related to population movements and partnership patterns which will be collected over the course of the trial will be used in a parallel set of secondary analyses to identify risk factors that influenced the risk of transmission. If the intervention does not have a significant impact, these important analyses will help to identify which factors were likely to have led to the trial result.
Data analysis and preparation of papers will be led by various members of the TasP  The trial co-ordinator/HIV physician will notify each SAE to the trial investigator at the Africa Centre as soon as he is aware of the event. The trial co-ordinator/HIV physician will complete the trial specific "Initial Serious Adverse Event Notification" form. The form will include: study ID of participant, date of SAE and detailed description of the SAE.
The trial investigator at the Africa Centre will sends the notification to the ANRS pharmacovigilance unit within two working days of being notified of the event.

Pregnancy reporting
Pregnancy is not considered as an AE or SAE. However the investigator must report all pregnancies occurring in HIV-infected women on ART during the trial on the "Pregnancy notification form" as soon as these are confirmed. Women of child bearing age will be asked about the possibility of being pregnant at each trial clinic visit; women who consider they may be pregnant will be offered an immediate pregnancy test As soon, as a pregnancy becomes known, the nurse in charge of the patient should immediately report to the physician to adapt ART as needed.
Any pregnancy occurring during the trial and its outcome should be immediately reported to ANRS, using the trial "pregnancy notification form". The participant has to agree with collecting data on her pregnancy.
The investigator has to notify each pregnancy as soon as the investigator is aware to coordinating centre (same system as for SAE above, we can just say see section 9.2).
The ANRS pharmacovigilance unit reports all pregnancies in the International Antiretroviral Pregnancy Registry.
The medical surveillance of the women and their children will be reinforced, specifically regarding serious pathology occurring during pregnancy and any congenital abnormalities in the infant at delivery. A SAE initial report form should be filled if any anomaly is detected. Figure 6 below summarises the trial oversight organisation.

Figure 6
Trial oversight

Steering Committee (SC) and Coordination Team (CT)
The Steering Committee (SC) will be responsible for the conduct of the trial and its overall organization. The SC is the trial decision body, for all scientific and administrative aspects. The SC will be co-chaired by the Coordinating Investigators. It will comprise the team investigators in South Africa, France and Switzerland and representatives of the Sponsor as listed on page 3 (The ANRS 12249 TasP Trial Team). The SC will meet as regularly as needed on conference calls and face-to-face meetings.
The SC ensures the correct implementation of the study and compliance with the protocol, and verifies its ethical compliance. It decides about any relevant changes to the protocol, necessary for continuation of the study.
The SC is responsible for the scientific promotion and communication of the trial data. Any sub-studies using the trial data should be discussed and approved by the SC.
The SC written report is sent to the investigators and to the Sponsor.
The Coordination Team (CT) is composed of the three Coordinating Investigators, the two Project Coordinators (one in South Africa and one in France), and any relevant participants to discuss specific issues. The CT undertakes the day-today management of the trial. The CT will meet by teleconference, which will usually be held at monthly Sponsor intervals although may be needed more frequently in the initial phases, frequency to be decided by the Project Coordinators.
The CT reports regularly to the SC for updates on trial progress and potential issues and, sends a quarterly progress reports (the content will be defined jointly at the beginning of the trial) to the Sponsor and to the Steering Committee.
The CT together with the SC coordinates and prepares the progress and scientific reports, communications and presentations to the SAB, DSMB, Ethical Committees and national regulatory bodies.

Data Safety Monitoring Board (DSMB)
The Data Safety Monitoring Board (DSMB) is the group that monitors the main safety and efficacy outcome measures and the overall conduct of the trial, with the aim of protecting the safety and the interests of the trial participants. The DSMB is appointed by the Sponsor before the implementation of the study and follows the current ANRS procedures "Independent Committee for ANRS sponsored clinical trials in developing countries".
The DSMB is an independent consultative committee in charge of alerting the scientific committee, the Coordinating Investigators and the Sponsor of any modification in the trial risk benefit ratio.
The DSMB is composed of six independent experts covering the main disciplines of the trial (biostatistics, HIV adult medicine, prevention, and bioethics). All members must be free from any direct involvement with the trial; any competing interests, both real and potential, must be declared.
The DSMB meets on a regular basis (frequency is decided by the board) throughout the trial and at least once a year. In the event of a serious or unexpected problem, an extraordinary meeting may be requested by the SC, the SAB Chair or the Sponsor to discuss on questions relative to the scientific and ethical integrity of the study. The first session of the DSMB will be held prior to the implementation of the trial to agree with the trial investigators the stopping rules needed to monitor the safety issue.
The DSMB, appointed for the needs of the study, will have access to all intermediate data and results decoded by cluster, as well as to any information justifying any change affecting the course of the study. It will monitor the trial regularly with a focus on issues relating to quality of trial conduct, such as rates of recruitment, adherence to trial interventions, visit schedules, losses to follow-up, respect of ethical principles and safety data.
DSMB meetings will be organized in two parts: an open session with the trial investigators and sponsor representatives followed by closed sessions with DSMB members and the trial statistician if relevant.
DSMB written reports signed by the chair will be sent to the Coordinating Investigators, the chair of the trial SAB, the chair of the Steering Committee and the sponsor.

Scientific Advisory Board (SAB)
The Scientific Advisory Board (SAB) is the group that oversees the overall conduct of the trial. Its mission is to make sure and to report, particularly to the Sponsor, whether the study is carried out properly scientifically, ethically and logistically.
The SAB is appointed by the Sponsor before the implementation of the study.
A first meeting will be convened prior to the beginning of the trial and the following meetings will be scheduled at least once a year till the end of the trial. In the event of a serious or unexpected problem, an extraordinary meeting may be requested by the SC, the SAB Chair or the Sponsor. Meetings of the SAB will be organized whenever possible in South Africa.
The SAB formulates written recommendations to the SC and to the Sponsor. The SC is expected to give its written feed back to the SAB and the Sponsor.

Ethics and competent authorities
This trial will be conducted in compliance with the protocol and with the following: This final version of this protocol was approved and signed by the three Coordinating Investigators and the Sponsor (cf. signatures on page 1).

Africa Centre Community Advisory Board (CAB)
The Africa Centre's Community Advisory Board (CAB) is an autonomous body consisting of approximately 30 members chosen by the community, whose main role is to link the Africa Centre with the community. The CAB membership comprises representatives from the traditional authorities, local councillors, community members and representative from the local offices of key provincial government departments including Health, Education and Social Development. Through the CAB, community input will be provided into trial design (cohort selection criteria), questionnaire content, participant follow-up plans, informed consent procedures, risk reduction interventions, community education and outreach, recruitment, retention planning and dissemination of trial findings. All research initiatives at the Africa Centre are presented to the CAB for discussion before seeking ethics approval from the University of KwaZulu-Natal's Ethics Committee.

Biomedical Research Ethics Committee of the University of KwaZulu-Natal
The Biomedical Research Ethics Committee is mandated to fulfil its function by the Senate of the University of KwaZulu-Natal. The essential function of the Committee is to review the protocols of all human subjects health research projects proposed to be undertaken by students and members of staff of the University. The purpose of this review is the protection of the dignity, rights, safety and well-being of all human participants of research. Special attention is given to research that may include vulnerable participants. The Committee is available to review, advise on, and approve or reject research protocols involving human participants submitted to it by researchers. Research to be reviewed will be in accordance with the stipulations of the National Health Act of the Republic of South Africa.
The membership of the Committee comprises of a Chairperson, two Deputy Chairs, at least two laypersons with no affiliations with the institution (preferably from the community), at least one member with knowledge of and current experience in research areas that are regularly considered by the Committee, at least two members with knowledge of and current experience in professional care, counselling or treatment (e.g. general practitioner, psychologist, etc.), at least one member who is legally trained.
Decisions are made at meetings at which at least a quorum of 50 % plus one is present.

KwaZulu-Natal Department of Health (DoH)
The mission of the KwaZulu-Natal Provincial DoH is to develop a sustainable, coordinated, integrated and comprehensive health system at all levels, based on the primary health care approach through the district health system.

Community, participant and patient information and consent
Community engagement and provision of adequate patient/participant information are keys in the success of the trial. The KwaZulu-Natal DoH has developed and approved an extensive array of community and patient information materials relating to all aspects of HIV -prevention, treatment and care. These are already widely available in the clinics and will be made available in the trial clinics as well. The social science team, with the community liaison office, is developing other information and community education materials for use in the trial as part of the initial preparatory work started in 2010. Examples of the materials being prepared include a community information leaflet (see Appendix 15.4), and the referral cards given to all participants during the homebased testing rounds. The four community advisory panels described above and Africa Centre Community Liaison Office will also contribute to the development of community, participant or patient information materials to ensure their appropriateness. Development and provision of any printed materials will follow normal approval processes in respect to their scientific content, input from trial Steering Committee, trial ethics and the provincial DoH as appropriate.
We will use the routine community road shows that are part of the Hlabisa Treatment and Care Programme's activities to ensure that there is a continuous feedback loop between the investigators and the communities. Over the years these road shows have been demonstrated to be a highly effective communication tools to promote HIV testing and treatment, and to educate the communities about the range of services and options provided by the programme. The content of any materials to be disseminated through the road shows will be approved by the trial investigators.
Consent for the collection of data and the retention of residual samples will be collected from individuals following the procedures of the Africa Centre and approved by the University of KwaZulu-Natal Biomedical Research Ethics Committee. Consent forms are presented in appendix 15.4. In each instance when consent is sought an individual will be fully briefed by the counsellor or clinical staff member and asked to provide a signature, the signature will then be countersigned by the staff member with the inclusion of a date time. For participants unable to write their own name, they will be asked to 'make their mark' (customarily an 'X'). In the event that a participant needs to do this the counsellor will need to have a second witness, normally another household member or Africa Centre counsellor, to verify the mark. This will only apply during the home-based testing rounds and when a person is not in possession of a South African national identity document. For participants that attend trial clinics for treatment and care, who are unable to write and do not possess a South African national identity document the process of making a mark/providing a thumb-print will be witnessed by two clinic staff members, unless the person attends with a 'treatment supporter' or family member, to avoid the potential for inadvertent disclosure of HIV status.

Data confidentiality
Current Africa Centre procedures require all staff to sign confidentiality and Acceptable Use Policy agreements, stating they have read and understood them and that they agree to abide by them. There are also formal policies in place regarding e.g. password usage and email security.
Each trial subject will be identified by a unique trial number and this alone will be used throughout the trial to identify the participant.
Personal identification details (Name, Id. No, clinic numbers, address, etc.) will only be made available to a) those whose job within the operational activities of the trial makes having such information absolutely essential, and b) to senior members of the trial administration (coordinating investigators, ANRS trial monitors, DSMB statistician) at the discretion of the trial coordinating investigators.
Completed questionnaires will be stored in a secure environment (in locked cabinets within the secure, access-controlled, Data Centre) and access will be granted only on a 'need to know' basis. These completed questionnaires and other study-related documents will be digitally archived according to the SOP developed in the AC for this purpose following strict quality control procedures.
Staff will be required (and trained) to always email personal information in passwordprotected attachments, not in the body of email, nor in unprotected attachments.
Similarly, transmission of personal information via text (SMS) or fax messages, even for valid operational purposes, will be forbidden.
Data used for analysis, as opposed to day-to-day operational activities, will never contain personal identifying information, and will only be issued under a formal Data Use Agreement which must be signed by the data user, the Africa Centre Director, and the trial coordinating investigators. It imposes conditions on data use including that data will be stored securely, will not be passed to others, used only for the agreed purposes, and that it will be destroyed after completion of the agreed analyses, as per Africa Centre standard procedures.

Protocol amendments
Any substantial change to the protocol will be described in an amendment, which will be ratified by the trial SC, and forwarded to the sponsor for agreement and to the pharmaceutical company partners in the trial (where appropriate) for information and comments. These amendments will then be forwarded to the Biomedical Research Ethics Committee of the University of KwaZulu-Natal for approval. Any amendment should be signed by the Sponsor, the Coordinating Investigators and the Biomedical Research Ethics Committee of the University of KwaZulu-Natal before its implementation.

Investigators responsibilities
The Coordinating Investigators agree to conduct this trial in full accordance with the provisions of this protocol and will comply with all requirements regarding the obligations of clinical investigators as outlined in the ICH -E6 good clinical practice (GCP) guidelines. They agree to maintain all trial documentation until the trial sponsor consents to disposal of files in writing. They are fully aware of the potential risks and side effects of the products under investigation, and will ensure that all associates, colleagues, and employees assisting in the conduct of the trial are informed about the obligations incurred by their contribution to the trial.

Insurance
The ANRS, as the Sponsor, will take a civil liability insurance policy, for the trial to cover all HIV infected patients receiving ARV treatment. A copy of the insurance certificate is given in Appendix 15.6.

Data collection and storage of forms
Questionnaires, electronic or paper-based CRFs will be used for data collection on site. The questionnaires and CRFs will be designed, completed, stored and relayed according to GCP principles and using local accredited GCP trainers. The completed questionnaires, CRFs and supporting documentation will be kept securely, in locked cabinets at the Africa Centre location, as trial source documents for potential review and/or audit during or after the trial. No source documents will be destroyed without specific permission in writing from the project coordinators. Only the coordinating investigators, the project coordinators and the authorised trial personnel will have access to the completed questionnaires, CRFs and supporting documents. Data capture and storage will be undertaken using computer systems compliant to GCP. To ensure correct operation according to SOP all system users will be trained and evaluated on a regular basis in line with Africa Centre policy. In-country and external monitoring will be organized to ensure that all the trial procedures are respected on site and to verify the data validity and reliability.

Data management
The Africa Centre Information Technology (IT) network is a technologically advanced Microsoft-based setup, professionally designed and maintained according to Microsoft 'best practices', by a large South African IT services company in collaboration with the Africa Centre IT Manager and his staff. Servers (mostly virtualised) are all located at the Africa Centre in a secure computer room and the network is protected by uninterrupted power supply (UPS) firewalls, up-to-date virus and malwares scanning software. Users are all given their own logins and sign confidentiality and 'Acceptable Use Policy' agreements. A comprehensive system of backups and archiving is in place, with some held off-site (transported and stored by a professional security company). A disaster recovery plan is in place and reviewed regularly.
The data collection, entry and management for the trial will be modelled closely on the current practices and procedures for existing Africa Centre surveys and trials for which a large suite of SOPs and policy documents exist (see www.africacentre.com) and which have been developed over more than 10 years of field operations. Many of the staff involved have considerable experience of managing clinical trials and the surrounding business processes as well as our surveillance systems and various demographic and health surveys. Currently the Data Centre prints and processes about 1.2 million pages per annum.
Every questionnaire will be individually bar-coded, with its details recorded in a database, and scanned at each stage of its life from printing and issue to fieldworkers through to final archiving, hence achieving an end-to-end 'chain of ownership' and clear visibility of where every form is at any moment. All completed forms will, on return from the field, first be checked visually by our Quality Control Department before being passed to staff in our secure Data Centre for data entry, and subsequent digital scanning and archiving. All entered data will be carefully checked for consistency and accuracy and, if necessary, correction.
All data will be stored in a MS-SQL Server database located on one of our own database servers, managed by professional Database Administrators. Access to read, enter, modify or delete data will be granted via the standard authentication and access-control features of MS-SQL Server and MS-Windows. Laboratory results will be transmitted, using already-established procedures, directly into the database, via a secure (https) connection from the Africa Centre Laboratory's Laboratory Information System. Data issued for analysis by scientists will all be anonymised and covered by formal, signed Data Use Agreements, which cover acceptable use, security, destruction after use etc.
Overall, the Africa Centre offers a very high quality, tried and tested professional data collection and management service, conforming to good clinical practice and trial specifications. Compliance with the ANRS rules and regulations will be verified before trial start.

Exchange of data
The trial data are stored at the Africa Centre. All members of the SC can have access to specific trial data as per their research interests, as appropriate and following agreed procedures. Requests for data for specific analyses must be accompanied by a data analysis plan, as well as, if required, ethics certification by recognised ethics committees abroad. All requests for analysis will need to be discussed and approved by the SC.
Any exchange of trial data will be through secure connections, password-protected (see trial SOPs for further details).

Quality assurance, monitoring, audits and inspections
All trial procedures will be standardised and compiled in a manual of operations. A training session will be scheduled before the trial starts for all trial staff.
In accordance with Good Clinical Practice (GCP) to ensure the quality of the trial, each investigator accepts monitoring visits, audits and inspections.

Monitoring
A Trial Initiation Visit will take place before the first patient enrolment in the trial. During this visit, the monitor/monitoring team (from UKZN clinical trials centre or from ANRS) will review the trial material: documents compiled in an Investigator file, trial products and will verify that the investigational team understands the protocol and GCP requirements.
During the trial, a representative of the Africa Centre/monitoring team will make regular Monitoring Visits to the clinical service, hospital pharmacy and laboratories involved in the trial to (i) ensure that the study is conducted according to the protocol and GCP and (ii) to help the investigational team in solving problems. During these monitoring visits, the monitor will have access to the source document for patient validation data reported in CRF (at any time, the investigator or his representative can be contacted for any matter relating to the protocol, its practical application or the measures to take facing certain events).
Once all participants have been enrolled, and all study procedures have been completed, and all clinical data has been duly recorded in the CRFs and reported to the Sponsor, a Close-out Visit must be done by the monitor, to ensure that the Investigator File and other trial documents are archived properly; in addition, the monitor must collect all unused trial material, documents and products.
The monitor will submit a written report to the sponsor or its representative after each trial-site visit or trial-related communication.

Audits and inspections
The trial may be audited by the Sponsor or with his express authorization by other agencies.
The purpose of a sponsor's audit, which is independent of and separate from routine monitoring or quality control functions, is to be able to evaluate trial conduct and compliance with the protocol, SOPs, GCP, and the applicable regulatory requirements.

Archiving
The investigators on site will digitally archive the patient information sheet, informed consent forms, protocol and amendment(s), CRF model, correspondences, patients' CRF and sources data/documents and will keep this information for 15 years at the Africa Centre, in accordance with GCP for clinical trials.

Results publications
The investigator team will meet to define the core publications from the first phase, responsibility for overseeing additional papers and plans for analysis (including the core papers) will be overseen by a publication committee.
The main trial results on the effectiveness of 'treatment as prevention' must receive prior approval from the coordinating investigators and SC in respect of the agreement contracted between ANRS and the pharmaceutical company, which provided drugs.
The publication of the main trial results should include the name of the sponsor, all the investigators who included or followed-up participants in the trial, the composition of the SC and the possible participation of pharmaceutical company, which provided drugs. Each publication in a scientific journal or subjected to a scientific conference or for the media should include: