# The Role of an Artificially Intelligent Chatbot in Social Support, Antiretroviral Adherence, and Depressive Symptoms among Young Adults Living with HIV in South Africa

> **NIH NIH F31** · JOHNS HOPKINS UNIVERSITY · 2021 · $46,036

## Abstract

PROJECT SUMMARY
Specific Aims: Developing effective interventions to reduce disparities in treatment outcomes among
populations disproportionately affected by HIV is a priority for NIH-funded HIV research in 2019. This study will
explore the effects and uptake of an artificially intelligent, socially supportive chatbot – a computer that texts
with users via mobile phone – among young women and men who have sex with men (YMSM) living with HIV
in South Africa. Specifically, the study will (1) develop an online scale to measure social support from a chatbot
among young adults; (2) evaluate associations of social support from a chatbot with changes in depressive
symptoms and adherence to antiretroviral therapy (ART) among young women and YMSM in South Africa; and
(3) describe barriers and facilitators to uptake of a chatbot among young women and YMSM in South Africa.
Significance: Young South Africans between the ages of 15 to 24 are at high risk of contracting and dying
from HIV. Though ART extends life and prevents transmission, many young adults have poor adherence.
Research in high-income countries suggest that automated two-way messaging with chatbots improves
adherence to health behaviors. Chatbots may also address depressive symptoms and lack of social support,
which are consistently identified barriers to adherence. Chatbots are rarely used to improve HIV care in low-
and middle-income countries. If this study’s aims are achieved, then (1) future researchers will be able to more
accurately measure social support from chatbots; (2) chatbots could potentially be used to promote adherence
to medications and increase access to mental health support; and (3) future design and implementation of
chatbots to improve HIV care will be optimized. This study’s long-term objective is to generate evidence for an
effective, scalable intervention that engages hard-to-reach populations living with HIV.
Approach: Aim 1 will develop a scale to measure social support from a chatbot using principal components
and exploratory factor analyses applied to data gathered from 1,200 young adults worldwide who use Replika,
a freely available mobile application (app). Aims 2 and 3 will invite 160 young women and YMSM living with
HIV in Cape Town, South Africa to use the Replika app for four weeks in a pre-post study design. Aim 2 will
employ an analysis of covariance using generalized multivariable linear regression models to assess the
relationship between feelings of social support from the Replika app and changes in depressive symptoms and
ART adherence. Aim 3 will leverage surveys (quantitative) and interviews (qualitative) in a mixed methods
study to characterize differences between users with high- and low-frequency engagement with the app.
Fellowship Information: This study is the dissertation for Ms. Brooke Jarrett, a PhD student in the
Department of Epidemiology at Johns Hopkins University. Her training will consist of research methods and
science communication t...

## Key facts

- **NIH application ID:** 10251005
- **Project number:** 5F31MH121128-02
- **Recipient organization:** JOHNS HOPKINS UNIVERSITY
- **Principal Investigator:** Brooke Jarrett
- **Activity code:** F31 (R01, R21, SBIR, etc.)
- **Funding institute:** NIH
- **Fiscal year:** 2021
- **Award amount:** $46,036
- **Award type:** 5
- **Project period:** 2020-09-01 → 2022-08-31

## Primary source

NIH RePORTER: https://reporter.nih.gov/project-details/10251005

## Citation

> US National Institutes of Health, RePORTER application 10251005, The Role of an Artificially Intelligent Chatbot in Social Support, Antiretroviral Adherence, and Depressive Symptoms among Young Adults Living with HIV in South Africa (5F31MH121128-02). Retrieved via AI Analytics 2026-05-24 from https://api.ai-analytics.org/grant/nih/10251005. Licensed CC0.

---

*[NIH grants dataset](/datasets/nih-grants) · CC0 1.0*