STUDY TITLE

Piloting an AI chatbot to improve patient experiences and increase engagement with HIV services in South African public health facilities: An implementation science study

SPONSOR: Gates Foundation

STUDY NUMBER: 35-2025

You are invited to take part in a research study that involves using an Artificial Intelligence (AI) chatbot on WhatsApp to guide you through various health-related topics, potentially including HIV and TB.

This document is to help you to decide if you would like to participate in this study. You should fully understand what is involved before you agree to take part. You should not agree to take part unless you are completely satisfied with the answers to your questions.

What is the purpose of this study?

The purpose of this study is to understand whether an artificial intelligence (AI) health chatbot called Coach Mpilo 1) meets users’ expectations and 2) improves users’ health-related knowledge, health-seeking behaviour, and health outcomes.

What is the Coach Mpilo chatbot?

The Coach Mpilo chatbot is a WhatsApp-based tool that can provide with users with information on health and wellbeing, including HIV, TB, and STIs, as well as guidance and support on how to manage those health conditions successfully and stay healthy.

Who is eligible to participate in this study?

Anyone over 18 years of age who has used the Coach Mpilo chatbot and can provide informed consent is eligible to participate.

What can the Coach Mpilo chatbot help me with?

  • Answering any questions you may have about health or wellness. The chatbot is particularly focused on HIV and TB, but you can ask it about anything you like.

  • Connecting you to other sources of information and support. For example, it can tell you where to find your nearest clinic or pharmacy, where to get an HIV self-test, etc.

  • Providing personalised reminders. If you like, it can give you reminders for when to take your medication, when to collect your refills, when to return to the clinic, etc.

  • Guiding you on self-testing. If you want to do an HIV self-test, the chatbot can guide you on how to do it properly. After the test, you have the option to upload a photo of the result for verification and get advice on treatment or prevention options.

What will I be expected to do in this study?

In this study, you will have the opportunity to interact with the Coach Mpilo chatbot on WhatsApp. Your interaction will be completely voluntary and on your own terms—you are not required to use it in any particular way or for any particular amount of time.

If you do use it, you may have the opportunity to provide feedback on the chatbot through brief surveys offered to you within the chat. These surveys will take less than one minute to complete, and you will not get more than one survey in a chat session. Surveys are completely voluntary, and you can choose to accept or decline, without it affecting your ability to continue chatting if you like.

What are my rights under this study?

Your participation in this study is completely voluntary. You can choose not to participate, and if you do choose to participate, you can change your mind and stop at any time without stating any reasons. Your decision not to participate or your withdrawal will not affect your access to health care or have any other negative consequences. If you choose to stop using the chatbot, you can stop at any time by selecting this option on the menu, without needing to notify anyone.

When a survey is offered within the chatbot, you have the right to decline to answer and can simply close the survey. Declining or ignoring the survey will not impact your chat experience in any way.

You have the right to request that your data be deleted from the chatbot, via an option in the main menu. If you request your data to be deleted, this will be done within 30 days of making the request.

What are the risks of participating in this study?

Some topics, like HIV and TB, may be sensitive, and using the chatbot could cause a small amount of emotional discomfort. The chatbot will be monitored for any situations where you may need additional support, and a real-life coach will then follow up with you. Again, you also have the right to stop your participation at any time.

We will aim to keep your identity and responses private. However, there is a small chance that someone could accidentally learn about your participation or the information you share. For example, if someone else uses your phone. We will provide you with tips on how to prevent this and show you how to hide or delete chats.

How will you protect my data and ensure my privacy?

Your data is covered by the Protection of Personal Information Act (POPI Act) and the study team is required to comply with the Act in all collection and handling of your data.

Your participation in this study is confidential and your participation will be anonymised in all analyses and reporting. Your identity will not be revealed while the study is being conducted or when the study findings are reported publicly. We will not record your full name anywhere and no one will be able to connect you to the answers you give. Your answers will be linked to a code number. You are also welcome to use a pseudonym (another name) during your chats.

You will not be required to share any personal information with the chatbot. You may be asked for information that allows the study team to match your use of the chatbot with your use of in-person health services, for analysis purposes only. Sharing of this information is completely voluntary, and anything you may decide to share or not share is entirely your decision. If you do share this information, the study team will replace all personal information with a unique code so that you cannot be identified.

Anything that you do share in your chats, including personal information, will remain confidential and will not be shared with anyone outside the study team. The number of the phone you use to chat with the chatbot will also be kept confidential and only accessible to members of the study team. Any data that are publicly reported will be combined with data from other users and analysed as a group rather than individually.

All identifying information about you will be kept in a locked file cabinet or a password-protected hard drive. This data will not be available to anyone who is not part of the study and will be kept confidential to the full extent possible by law.

Your data will be kept for five years and thereafter will be destroyed. Paper records will be shredded and recycled. Records stored on a computer hard drive will be erased using software applications designed to remove all data from the storage devices. A record showing which data was destroyed, and when and how it was destroyed will be kept in a locked file cabinet.

Data collected during the study will only be used for research purposes. Findings from study might be shared publicly, but this will be done without revealing the identity of any study participant.

Records from your participation may be reviewed by people responsible for making sure that the study is done properly. All of these people are required to keep your identity confidential and have signed agreements to do so. Records that identify you will be available only to people working on the study unless you give permission for other people to see the records.

For your own benefit, your chats may also be reviewed and followed up by a member of the study team if there are any indications of risk of self-harm, harm to or from others, or other situations where additional support may be needed. However, you will still have the option to respond or not respond to any follow-up messages.

What are the benefits of participating in this study?

By participating, you will have access to private, confidential health information and support through the chatbot. The findings from this study may also benefit you and members of your community by helping the study team to improve the chatbot for other users in the future. Your contribution might, in a small way, go towards benefiting the community.

Will I be reimbursed for participating in this study?

There is no compensation for your interaction with the chatbot or for completion of brief surveys within the chat.

Who has reviewed and approved this study?

The protocol for this study was submitted to the Foundation for Professional Development Research Ethics Committee (FPD-REC), a research ethics committee registered with the National Health Research Ethics Council. Written approval has been granted by FPD-REC for the conduct of the study.

Who can I contact if I have questions or concerns about this study?

If you have any questions or concerns about this study, you may contact the Principal Investigator, Mr. Shawn Malone, Principal Investigator, smalone@psi.org, +27 81 038 1862.

If you have any questions or concerns regarding the ethical conduct of this study, you are encouraged to contact the FPDREC, Liviav@foundation.co.za, +27 87 821 1109.