“Your Choice” – Using artificial intelligence (AI) to reduce stigma & improve precision in HIV risk assessment and reporting of results

Partner organizations: Audere and Chaps

 

 

Study objectives: In this study, we co-designed with at-risk populations and validated a prototype of the “Your Choice” app, utilising multiple Large Language Models (LLMs) to guide empathetic conversations to gather behavioural and needs data related to HIV vulnerability for clients and provide a potential, tailored HIV prevention plan.

 

Objective 1: Assess the usability, acceptability and appropriateness of the “Your Choice” application for HIV vulnerability screening among recipients of care.


Objective 2:
Assess the feasibility and agreeability of the information from the “Your Choice” app among health care providers.

 

 

Intervention description: The “Your Choice” app, built on the HealthPulse platform, is an innovative collaboration between Audere, a leading digital  health AI nonprofit organization, CHAPS, a prominent organisation dedicated to promoting sexual health and Indlela. Leveraging the power of Large Language Models LLMs, this app aims to revolutionise healthcare by serving as a crucial intermediary between healthcare providers and individuals seeking care. Specifically, the app focuses on efficiency for in-clinic screening of community members for HIV vulnerability, helping healthcare professionals make informed decisions about needs and conversations about Pre-Exposure Prophylaxis (PrEP) interest and fit. By utilising advanced algorithms and up-to-date medical information, the app will streamline the process, ensuring that those who can benefit most from PrEP receive the support they need. For this study the app was in the prototype phase and was not used in a clinical encounter nor was it used for any clinical decision making.

 

 

Evaluation description: We utilised a mixed method design to assess the usability, acceptability and appropriateness of the “Your Choice” application for HIV vulnerability screening among community members. We also assessed the feasibility and agreeability of the information from the “Your Choice” app among health care providers.

 

 

Study outcomes: To measure success in the current formative project we used a combination of quantitative and qualitative approaches.

 

Quantitative measurements

Usability, acceptability and appropriateness of the “Your Choice” application for HIV vulnerability screening and interest among community members and recipients of care.

 

Recipients of care:

Acceptability of intervention measure (AIM)

Intervention appropriateness measure (IAM)

Usability using the system usability scale (SUS)

 

Health care providers:

Feasibility of intervention measure (FIM).

Usability using the system usability scale (SUS)

 

Qualitative measurements

We conducted qualitative interviews on the app experience with 35 recipients of care and with 20 health care providers. For each group we considered questions in the following domains:

 

Recipients of care: Experience using the app for HIV vulnerability conversations, suggested areas of improvement, challenges today, what they liked most and implementation ideas.

 

Healthcare providers: Current HIV screening experience and shortcomings or challenges, experience with the app and the conversation summary provided, implementation challenges.

 

These data contributed to acceptability, usability and system usability.

 

 

Study site: The study was conducted at the Indlela offices in Johannesburg with recipients of care and at clinics in Gauteng with healthcare providers.

 

Target population:

Objective 1 (Recipients of Care)

 

Inclusion criteria 

  • HIV negative by self-report at B-Hub enrolment
  • Willing to provide written informed consent
  • Comfortable consenting and talking in English
  • ≥18 years old

 

Exclusion criteria

  • Participants who are not willing or available to travel to the study site to take part in this sub-study

 

Objective 2 (Healthcare providers)

Inclusion criteria

  • Willing to provide written informed consent
  • Comfortable consenting and talking in English
  • Licensed to prescribe PrEP
  • Prescribed PrEP in the last 6 months

 

Exclusion criteria

  • Participants who are not willing or available to take part in this sub-study

 

 

Sample size:

Objective 1: We enrolled up to 130 eligible B-Hub participants who are recipients of care and community members.

 

Objective 2: We engaged up to 20 B-Hub participants who are healthcare providers trained and engaged in providing health care services that include PrEP.

 

 

Duration: 7 Months