Choice-Based Support for Adults Starting or Restarting Antiretroviral Therapy in Cape Town, South Africa

Summary

Disengagement from HIV care is very common in the first year after starting or restarting antiretroviral therapy (ART). There is increasing recognition of people cycling in and out of HIV care over time people newly starting and restarting treatment after an interruption are at high risk of subsequent disengagement from care. While guidelines advocate for patient-centered models of care, patient preferences are often not considered fully in the design of interventions. Building on existing support intervention modalities, formative qualitative research and a stated preference survey, the investigators have designed a choice-based ART support intervention. The intervention offers a choice of a) in-person group support, b) low-touch WhatsApp support group, and c) individual digital support through the AI Coach chatbot.

This study will explore the feasibility, acceptability and preliminary impact of choice-based ART support for adults starting and restarting ART in Cape Town, South Africa, through a randomized pilot feasibility trial.

The objectives of this study are:

1. To determine the feasibility, acceptability, appropriateness and fidelity of the Zikhethele intervention components, including offering patients a choice of ART support intervention.
2. To describe the distribution of actual choices in the choice arm.
3. To describe outcomes in each pilot trial arm and explore the preliminary impact of offering a choice, compared allocation to a support intervention or standard of care, on patient empowerment and treatment outcomes.
4. To explore the hypothesised mechanisms of action and contextual moderators through in-depth interviews with participants and providers including consideration of patient empowerment, stigma and social support, and trust in provider, peers and digital tools.

A total of 140 adults starting or restarting ART will be consecutively recruited and randomised to a) standard of care (n=35), b) in-person support (n=35) and c) a choice (n=70) of in-person support, WhatsApp group support, AI coach, or no additional support (standard of care).

The intervention components will run for the first four months after start or restart, through to the first viral load and eligibility assessment for routine differentiated models of care. Briefly, the in-person group support will consist of monthly informal and discussion-based sessions framed around chronic medication adherence (including HIV, diabetes, and hypertension), designed to create a safe and supportive space where participants can share experiences, problem-solve, and build motivation to remain in care. The WhatsApp group support will be a virtual adaptation of the in-person model, designed to provide an accessible, low-barrier option for participants who prefer remote or flexible engagement. The AI Coach is a pilot AI chatbot (developed by Audere, PSI, HSRC, and Matchboxology, and being piloted in Gauteng and KwaZulu Natal by HE2RO at the Wits Health Consortium) inspired by the in-person Coach Mpilo model-a peer navigator case management approach in South Africa that employs men living with HIV as "coaches" to support linkage, retention, and re-engagement in care. The AI coach is available anytime via WhatsApp, and offers trusted information (through a curated large-language model), and empathetic counselling and behavioural nudges to encourage healthy habits.

Individuals (aged 18 and older, living with HIV and currently attending the clinic to start ART for the first time or to restart after an interruption of 3 or more months) will be approached during their routine clinic visit by a trained research fieldworker. Eligible individuals will undergo informed consent and be enrolled. Participants will complete an interviewer administered questionnaire at enrolment and at four months. Outcome data will also be abstracted from paper and electronic medical records, as well study and intervention logs, intervention debriefings and chatbot logs. A subset of 20 participants, purposefully selected to include each intervention component, will be invited to complete an in-depth interview at 4 months. Providers (n=6), while not involved directly in this pilot implementation, will also be invited to participate in an in-depth interview to explore the potential of this intervention.

This exploratory pilot study serves as a proof-of-concept for offering people living with HIV different modalities of support for engagement in care in the first four months. The study is not powered for efficacy but will provide valuable insights into feasibility (of both the trial design and intervention components) and acceptability, as well as real-world preferences and trade-offs. The investigators hypothesize that those able to choose a support intervention most appealing or most fitting to their life circumstances may have improved health empowerment which may in turn improve health outcomes.

Conditions

• HIV Care Loss to Followup

Interventions

BEHAVIORAL

In-person group support

The in-person group is framed around chronic medication adherence (including HIV, diabetes, and hypertension). It will consist of monthly informal and discussion-based sessions (60-90 minutes) designed to create a safe and supportive space where participants can share experiences, problem-solve together, and build motivation to remain in care. The group seeks to provide a semi-structured, peer-supported environment. Groups will consist of maximum 15 people. Participants will not be provided with any transport costs or reimbursement as this would not be feasible to scale. Tea and coffee and light refreshments will be provided. Each participant will be able to attend a minimum of three sessions over a four-month period. Sessions will be supported by a trained facilitator (peer living with a chronic health condition).

BEHAVIORAL

WhatsApp group support

The WhatsApp group support will be a virtual adaptation of the in-person group support model, designed to provide an accessible, low-barrier option for participants who prefer remote or flexible engagement. The group is framed as a virtual support community for people living with chronic health conditions such as HIV, diabetes, or hypertension. The group size will be 15-20 people. This virtual space aims to maintain the benefits of peer connection and mutual motivation while minimizing logistical and social barriers to participation. It builds on growing evidence for mHealth peer support in ART adherence and retention interventions in South Africa. Participants will be added to a single WhatsApp group called Zikhethele, moderated by a trained community health worker. The group will remain active for approximately four months, mirroring the duration of the in-person group. Members may join or leave the group at any time.

BEHAVIORAL

AI coach

AI coach (developed by Population Services International (PSI), Wits Health Consortium (WCI), and Audere) is a digital version of the in-person Coach Mpilo model where a peer navigator case management approach in South Africa that employs men living with HIV as "coaches" to support linkage, retention, and re-engagement in care. The digital version is an AI-based chatbot over WhatsApp, which aims to replicate the motivational, supportive functions of a human coach. The AI coach is available anytime, anywhere via WhatsApp, and offers trusted information (through a curated large-language model), and empathetic counselling and behavioural nudges to encourage healthy habits. Users are able to engage with the chatbot on their own phone or other device, via WhatsApp, at any time, from any location, free of charge (other than the use of WhatsApp data). The chatbot is not structured or scripted but rather designed to respond to any question or statement posed by the user.

BEHAVIORAL

Choice of support

Participants will receive a brief description of the available support options (in-person club, WhatsApp group, or AI coach) following a standard script. They will be able to discuss and select the modality that best fits their needs and circumstances. Participants in the choice arm will be asked to select one of the available intervention options, or they will be able to choose not to sign up for additional ART support (standard of care). The choice process will be facilitated by a trained community health worker, who will also support coordination and enrolment into the selected option. Offering choice has the potential to promote patient empowerment, strengthen engagement in care, and optimize treatment outcomes. Allowing patients to select a preferred health intervention, this can increase satisfaction, self-efficacy and health care empowerment, which in turn may improve engagement and health outcomes.

Sponsors & Collaborators

• Fogarty International Center of the National Institute of Health

  collaborator NIH

• National Institute of Mental Health (NIMH)

  collaborator NIH

• University of Cape Town

  lead OTHER

Study Design

Allocation

RANDOMIZED

Purpose

SUPPORTIVE_CARE

Masking

NONE

Model

PARALLEL

Eligibility

Min Age

18 Years

Sex

ALL

Healthy Volunteers

No

Timeline & Regulatory

Start

2026-02-28

Primary Completion

2026-08-31

Completion

2026-08-31

Countries

• South Africa

Study Locations