HIV Self-Testing Africa (STAR) Malawi: General Population

The aim of this project is to investigate the incremental costs and health benefits of adding HIVST to the remit of existing cadres of community volunteers in the general population in Malawi.

Specific objectives are to:

1. Validate the use of the OraQuick HIVST product in rural settings, establish preferences and social harms reporting systems, and conduct baseline surveys.
2. Carry out a pragmatic cluster-randomised trial with ART clinics and their catchment populations, including CBDAs, as the unit of randomisation.
3. Evaluate the impact of adding HIVST into the remit of VMMC mobilisers on demand for VMMC
4. Establish the expected costs and benefits of introducing HIVST to Malawi through economic and mathematical modelling.
5. Conduct interviews with policy-makers to prepare for national scale-up of HIVST and support the market introduction of quality-assured HIVST products.

CRCT METHODOLOGY/STUDY DESIGN

First stage randomization: HIVST v. SOC. The main study will be a cluster-randomised trial in up to 5 high HIV prevalence districts (provisionally Blantyre, Machinga, Mwanza, Neno and Thyolo) that already have community-based reproductive health services provided by PSI under funding provided by the German Technical Cooperation Agency (GTZ). This programme supports volunteers to socially-market reproductive and child health products in villages, with products stored and managed by collaborating primary care clinics. The unit of randomisation is the Primary Care + ART Clinic and surrounding catchment area villages to clinic. CBDAs in the HIVST Arm villages will be trained to provide HIVST services as well as reproductive health services. CBDAs in the SOC Arm villages will remain with reproductive health services only. Two separate units of evaluation will be pre-defined before randomisation: one or two evaluation villages that will have baseline-final household surveys, and a wider catchment area including all villages with CBDA activities.

Second stage randomization: HIVST with and without home assessment. Villages in the HIVST intervention area will be randomised to either HIVST only, or to HIVST plus offer of home assessment and HIV care initiation (first assessment and first 14 days of HIV care medications) with this additional intervention aimed at facilitating linkage into care.

CRCT SAMPLE SIZE

This trial has one primary and one secondary outcome, with a number of other analyses to be specified in the Statistical Analysis Plan. The sample size considerations here relate to the main Primary Outcome (comparison between arms of recent HIV testing).

A survey sample of 250 to 500 adults per cluster will provide sufficient power for a two sample comparison of unmatched proportions for HTC coverage was performed across the intervention and control arms to determine the number of clusters per arm.

The cluster size is based on the typical size of a rural village (250-500 people) from previous experience in Malawi. Using 2010 DHS data, baseline rates for individuals tested in the last 12 months is estimated at 25% to 40%, with a 45% to 60% predicted effect size in the intervention arm compared to the control arm. Baseline coverage for individuals who have ever tested is estimated at 42% to 60% with predicted effect size 30%-45%.

Figures 10 and 11 in the full protocol contain a table of scenarios for each of the outlined assumptions. To detect a 50% increase in rate of people testing in the past 12 months and a 45% increase in the percentage of people who have ever tested in the intervention villages, there should be approximately 8 clusters per arm and approximately 4,000 participants total. This study will aim to reach 10 clusters per arm and 5,000 participants to provide contingency for a lower than anticipated effect size.

FORMATIVE RESEARCH

Formative research will be conducted before the beginning of the cRCT, including (1) cognitive interviews to assess the suitability existing IFUs in this population and modify IFUs as needed; (2) key informant interviews to help identify needs of stakeholders working in this area; (3) community mapping to identify local services.

Cognitive interviewing to validate IFUS. Convenience sampling will be used to recruit clients from VCT clinics for the cognitive interviews. Eligibility criteria include clients who are 16 years or older, and willing to provide written informed consent for participation in the study. A witnessed thumbprint will be sufficient for those who are unable to read or write. Participants will be purposively selected to ensure a range in location and literacy level. Four to 12 participants will be recruited for each iteration of the cognitive interviews, with a maximum of 12 iterations of adaptation and trial of the intended IFUs (n = 144). Cognitive interviewing and iterative adaptation will continue until saturation occurs and no further modifications to user materials are made.

Key informant interviews. We will conduct key informant interviews to inform our understanding of needs through key informant interviews with stakeholders working in HIV policy, product regulation and HIV implementation. Approximately 45 participants will be selected purposively through snowball sampling for the KIIs. MLW-LSHTM, using their professional networks and PSI country staff, will identify appropriate stakeholders. Interviews will be with written informed consent, will use semi-structured questionnaires and will be audiotaped.

Community mapping. Primary care clinics and evaluation villagers will be selected in consultation with the program lead for PSI's CBDA programme, and in consultation with the relevant District Health Offices, Primary care clinics and the CBDAs themselves.Once villages have been selected, preliminary mapping exercises will be initiated including demarcation of the village boundaries and location of dwellings with geographical position system (GPS) and a brief household-level questionnaire (name of household head, demographics of household members).

PILOT STUDY

Two villages not included in the main study will be selected from rural Blantyre or Thyolo for pilot studies of accuracy and acceptability. From these villages, 200 to 250 participants in total will be recruited from either randomly selected households (n = 150-200) or community peer groups (n = 48) for the accuracy and acceptability studies .Participants who opt to self-test as well as undergo standard HTC will be given a HIVST kit and a test results forms to record their own reading of the results. They will then be provided with a demonstration and instructions before carrying out the self-test procedure in a private room and completing the test results forms. Participants who themselves recognise that they have made a user error or find that their first result is uninterpretable will be given a second test kit to repeat the procedure if they wish to do so.

Following self-testing, trained HTC counsellors will record their own reading of the test kit before completing the standard HTC process (parallel Unigold and Determine finger prick blood) and entering the results of blood-based testing. All patients testing HIV+ve will be referred to HIV care services.

PILOT STUDY SAMPLE SIZE

For the pilot study, 150 to 200 adults will be recruited from a random selection of 60 to 80 households (with an estimated 2-5 adults per household) using randomly selected GPS waypoints. All household members aged 16 years or older will be asked for written consent to answer a brief questionnaire (socio demographics, past HIV testing experience, chronic care including ART) and will be offered HIVST followed immediately by confirmatory HTC. Parental consent will be taken for 16 and 17-year-olds and witnessed thumbprint for individuals unable to read or write.